Synergistic action of polyanionic and non-polar cofactors in fibrillation of human islet amyloid polypeptide

Recent studies have led us to suppose that synergistic action of multiple solute cofactors could play substantial roles in amyloid-type fibrillogenesis of pathogenic polypeptides. To support this view, we performed aggregation experiments of human islet amyloid polypeptide (IAPP) in media containing both polyanions and non-polar solvents. The results demonstrated that the fibrillation at sub-micromolar IAPP occurred only when polyanionic and non-polar solutes coexist. A simple sum of two independent cofactor's effects could not account for the synergistic action. We propose that this synergy of polyanionic and nonpolar milieus could substantially modify the amyloidgenesis in the human body.     

P-protein of Chandipura virus is an N-protein-specific chaperone that acts at the nucleation stage

The nucleocapsid protein N of Chandipura virus is prone to aggregation in vitro. We have shown that this aggregation occurs in two phases in a nucleation-dependent manner. Electron microscopy suggests that the aggregated state may have a ring-like structure. Using a GFP fusion, we have shown that the N-protein also aggregates in vivo. The P-protein suppresses the N-protein aggregation efficiently, both in vitro and in vivo. Increased lag phase in the presence of the P-protein suggests that chaperone-like action of the P-protein occurs before the nucleation event. The P-protein, however, does not exert any chaperone-like action against other proteins, suggesting that it binds to the N-protein specifically. Surface plasmon resonance and fluorescence enhancement indeed suggest that the P-protein binds tightly to the native N-protein. The P-protein is thus an N-protein-specific chaperone which inhibits the nucleation phase of N-protein aggregation, thus keeping a pool of encapsidation-competent N-protein for viral maturation.     

Covalent blocking of fibril formation and aggregation of intracellular amyloidgenic proteins by transglutaminase-catalyzed intramolecular cross-linking

Two different types of physical bonding have been proposed to involve in the formation of neuronal inclusions of patients with neurodegenerative diseases such as Alzheimer's, Parkinson's, and polyglutamine diseases. One is the noncovalent bonding that stabilizes the amyloid-type fibrous aggregates, and the other is the covalent cross-linking catalyzed by tissue transglutaminase. The cross-linking is subdivided into the inter- and intramolecular cross-linking. Little attention has been paid to the pathological roles of the intramolecular cross-linking. To elucidate the possible interplay between the intramolecular cross-linking and the amyloid-type fibril formation, we performed an in vitro aggregation analysis of three intracellular amyloidgenic proteins (a domain of tau protein, alpha-synuclein, and truncated yeast prion Sup35) in the presence of tissue transglutaminase. The analysis was performed in low concentrations of the proteins using techniques including thioflavin T binding and mass spectrometry. The results demonstrated that the amyloid-type fibril formation was strongly inhibited by the transglutaminase-catalyzed intramolecular cross-linking, which blocked both the nucleation and the fiber extension steps of the amyloid formation. Far-UV CD spectroscopy indicated that the cross-linking slightly altered the backbone conformation of the proteins. It is likely that conformational restriction imposed by the intramolecular cross-links has impaired the ordered assembly of the amyloidgenic proteins. Nonamyloid type aggregation was also suppressed by the intramolecular cross-links. On the basis of the results, we proposed that tissue transglutaminase is a modulator for the protein aggregation and can act defensively against the fibril deposition in neurons.     

Inactivation mechanism of the beta-ketoacyl-[acyl carrier protein] reductase of bacterial type-II fatty acid synthase by epigallocatechin gallate.

Epigallocatechin gallate (EGCG), a major compound from green tea, reversibly inhibits beta-ketoacyl-[acyl carrier protein] reductase (FabG) from Escherichia coli. In this study, we found that EGCG exhibited an atypical time-dependent inhibition of FabG, which possibly resulted from the EGCG-induced aggregation of FabG. It was observed that FabG inactivation and aggregation occurred nearly simultaneously, with a lag time that decreased with increasing EGCG concentration. These results suggest that some chemical reactions, required for aggregation and inactivation, occurred during the lag time. Since EGC was detected by HPLC after the incubation of EGCG with FabG, EGCG probably covalently modified FabG. These further results showed that 1 tetramer of FabG must be modified by several, possibly 4, EGCG molecules before the formation of FabG aggregates. FabG aggregation was a first-order reaction independent of protein concentration. Due to an initial lag time, the first-order rate of aggregation gradually increased, reaching a maximal and constant value. The effect of increasing concentration of EGCG on the first-order rate constant for aggregation indicated that EGCG bound to FabG by affinity labeling. Based on the results, we propose a mechanism for the interaction of EGCG with FabG:EGCG first binds reversibly to each subunit of FabG, followed by covalent modification and then aggregation of the 4 EGCG-modified subunits.     

Cyclin A- and cyclin B-dependent protein kinases are regulated by different mechanisms in Xenopus egg extracts.

Cyclins are proteins which are synthesized and degraded in a cell cycle-dependent fashion and form integral regulatory subunits of protein kinase complexes involved in the regulation of the cell cycle. The best known catalytic subunit of a cyclin-dependent protein kinase complex is p34cdc2. In the cell, cyclins A and B are synthesized at different stages of the cell cycle and induce protein kinase activation with different kinetics. The kinetics of activation can be reproduced and studied in extracts of Xenopus eggs to which bacterially produced cyclins are added. In this paper we report that in egg extracts, both cyclin A and cyclin B associate with and activate the same catalytic subunit, p34cdc2. In addition, cyclin A binds a less abundant p33 protein kinase related to p34cdc2, the product of the cdk2/Eg1 gene. When complexed to cyclin B, p34cdc2 is subject to transient inhibition by tyrosine phosphorylation, producing a lag between the addition of cyclin and kinase activation. In contrast, p34cdc2 is only weakly tyrosine phosphorylated when bound to cyclin A and activates rapidly. This finding shows that a given kinase catalytic subunit can be regulated in a different manner depending on the nature of the regulatory subunit to which it binds. Tyrosine phosphorylation of p34cdc2 when complexed to cyclin B provides an inhibitory check on the activation of the M phase inducing protein kinase, allowing the coupling of processes such as DNA replication to the onset of metaphase. Our results suggest that, at least in the early Xenopus embryo, cyclin A-dependent protein kinases may not be subject to this checkpoint and are regulated primarily at the level of cyclin translation.     

Homotypic leukocyte aggregation triggered by a monoclonal antibody specific for a novel epitope expressed by the integrin β1 subunit: Conversion of nonresponsive cells by transfecting human integrin α4 subunit cDNA

The monoclonal antibody 33B6 was found to be specific for the β1 integrin subunit. Treatment of leukocytes with this antibody induced a vigorous homotypic aggregation that had similar physiologic conditions as aggregation induced by a monoclonal antibody specific for the α4 subunit. Expression of a β1 subunit on the cell surface was not sufficient for mAb 33B6-mediated aggregation to occur, since cells of the K562 erythroleukemia line failed to respond even though they expressed the β1 subunit and the 33B6 epitope. However, after transfection with cDNA encoding the α4 subunit, K562 cells acquired the ability to aggregate in response to mAb 33B6 binding. By contrast, mAb 33B6 blocked cell binding to the endothelial surface protein vascular cell adhesion molecule-1 and the extracellular matrix protein fibronectin. These results suggest that the β1 epitope defined by mAb 33B6 may play a novel role in regulating leukocyte adhesive interactions.     

Amyloid β Oligomeric Species Present in the Lag Phase of Amyloid Formation

Alzheimer’s disease (AD)-associated amyloid β peptide (Aβ) is one of the main actors in AD pathogenesis. Aβ is characterized by its high tendency to self-associate, leading to the generation of oligomers and amyloid fibrils. The elucidation of pathways and intermediates is crucial for the understanding of protein assembly mechanisms in general and in conjunction with neurodegenerative diseases, e.g., for the identification of new therapeutic targets. Our study focused on Aβ42 and its oligomeric assemblies in the lag phase of amyloid formation, as studied by sedimentation velocity (SV) centrifugation. The assembly state of Aβ during the lag phase, the time required by an Aβ solution to reach the exponential growth phase of aggregation, was characterized by a dominant monomer fraction below 1 S and a population of oligomeric species between 4 and 16 S. From the oligomer population, two major species close to a 12-mer and an 18-mer with a globular shape were identified. The recurrence of these two species at different initial concentrations and experimental conditions as the smallest assemblies present in solution supports the existence of distinct, energetically favored assemblies in solution. The sizes of the two species suggest an Aβ42 aggregation pathway that is based on a basic hexameric building block. The study demonstrates the potential of SV analysis for the evaluation of protein aggregation pathways.     

Inhibition of cdc2 activation by INH/PP2A.

INH, a type 2A protein phosphatase (PP2A), negatively regulates entry into M phase and the cyclin B-dependent activation of cdc2 in Xenopus extracts. INH appears to be central to the mechanism of the trigger for mitotic initiation, as it prevents the premature activation of cdc2. We first show that INH is a conventional form of PP2A with a B alpha regulatory subunit. We next explore the mechanism by which it inhibits cdc2 activation by examining the effect of purified PP2A on the reaction pathways controlling cdc2 activity. Our results suggest that although PP2A inhibits the switch in tyrosine kinase and tyrosine phosphatase activities accompanying mitosis, this switch is a consequence of the inhibition of some other rate-limiting event. In the preactivation phase, PP2A inhibits the pathway leading to T161 phosphorylation, suggesting that this activity may be one of the rate-limiting events for transition. However, our results also suggest that the accumulation of active cdc2/cyclin complexes during the lag is only one of the events required for triggering entry into mitosis.     

Kinetic studies of protein L aggregation and disaggregation

We have investigated the aggregation of protein L in 25% (vol/vol) TFE and 10 mM HCl. Under both conditions, aggregates adopt a fibrillar structure and bind dyes Congo Red and Thioflavin T consistent with the presence of amyloid fibrils. The kinetics of aggregation in 25% TFE suggest a linear-elongation mechanism with critical nucleus size of either two or three monomers. Aggregation kinetics in 10 mM HCl show a prolonged lag phase prior to a rapid increase in aggregation. The lag phase is time-dependent, but the time dependence can be eliminated by the addition of pre-formed seeds. Disaggregation studies show that for aggregates formed in TFE, aggregate stability is a strong function of aggregate age. For example, after 200 min of aggregation, 40% of the aggregation reaction is irreversible, while after 3 days over 60% is irreversible. When the final concentration of the denaturant, TFE, is reduced from 5% to 0, the amount of reversible aggregation doubles. Disaggregation studies of aggregates formed in TFE and 10 mM HCl reveal a complicated effect of pH on aggregate stability.     

Sequence of critical events involved in fusion of phospholipid vesicles induced by clathrin.

Membrane fusion induced by clathrin is accompanied by several events such as conformational change, membrane binding and association of clathrin, and membrane aggregation (Maezawa et al. (1989) Biochemistry 28, 1422-1428; Maezawa and Yoshimura (1990) Biochem. Biophys. Res. Commun. 173, 134-140). To clarify the sequence of these events, we examined their time-courses by reducing the pH of the medium from 7.4 to a given pH in the range of 3.5-5.0 at 25 degrees C or 10 degrees C. Large unilamellar vesicles composed of phosphatidylserine and phosphatidylcholine were used in most experiments. The half-time for conformational change of clathrin was less than those for membrane binding and association of clathrin. The half-times and the initial rates of membrane binding and association of clathrin were similar order of magnitude, although the pH-profiles of the initial rates of the two events were somewhat different. Membrane aggregation started after membrane binding of clathrin. A lag phase was observed in the time-course of membrane fusion, whereas there was no lag phase in membrane binding and association of clathrin and membrane aggregation. Moreover, the lag time before fusion was independent of the clathrin concentration, although the initial rates of these three events were dependent on it, suggesting that the three reactions are not responsible for the lag phase before fusion, and that there is some other event(s) in the lag time. On the other hand, there was a threshould-pH in the pH profile of the lag-time and the threshold-pH coincided with the critical pH at which the final associated state of clathrin was apparently reversed in the presence and absence of liposomes, suggesting that the event(s) in the lag phase may be related to this final associated state of clathrin molecules on the liposome membranes. These results indicate that clathrin-induced fusion of liposomes is initiated through the following sequential events: conformational change of clathrin, membrane binding and association of clathrin, which occur simultaneously but independently, membrane aggregation, an event(s) in the lag phase, and actual fusion.     

DnaK prevents human insulin amyloid fiber formation on hydrophobic surfaces

We have developed a multiwell-based protein aggregation assay to study the kinetics of insulin adsorption and aggregation on hydrophobic surfaces and to investigate the molecular mechanisms involved. Protein-surface interaction progresses in two phases: (1) a lag phase during which proteins adsorb and prefibrillar aggregates form on the material surface and (2) a growth phase during which amyloid fibers form and then are progressively released into solution. We studied the effect of three bacterial chaperones, DnaK, DnaJ, and ClpB, on insulin aggregation kinetics. In the presence of ATP, the simultaneous presence of DnaK, DnaJ, and ClpB allows good protection of insulin against aggregation. In the absence of ATP, DnaK alone is able to prevent insulin aggregation. Furthermore, DnaK binds to insulin adsorbed on hydrophobic surfaces. This process is slowed in the presence of ATP and can be enhanced by the cochaperone DnaJ. The peptide LVEALYL, derived from the insulin B chain, is known to promote fast aggregation in a concentration- and pH-dependent manner in solution. We show that it also shortens the lag phase for insulin aggregation on hydrophobic surfaces. As this peptide is also a known DnaK substrate, our data indicate that the peptide and the chaperone might compete for a common site during the process of insulin aggregation on hydrophobic surfaces.     

Activity of covalently cross-linked cholera toxin with the adenylate cyclase of intact and lysed pigeon erythrocytes.

Reaction of cholera toxin with NN'-bis(carboximidomethyl)tartaramide dimethyl ester produced several cross-linked species that had subunit B (which binds to the cell surface) and peptides A1 (which activates adenylate cyclase) and A2 all covalently joined together. This cross-linded material had activity with pigeon erythrocytes that was comparable in all respects with that of native toxin. It activated the adenylate cyclase of whole cells, showing a characteristic lag phase, and this activation was increased if the cells had been preincubated with ganglioside GM1, but abolished if the protein had been preincubated with the ganglioside. It activated the enzyme in lysed cells more strongly and without the lag phase. These results show that the toxin is active even when peptide A1 cannot be released from the rest of the molecule.     

Amyloid fibril formation of hen lysozyme depends on the instability of the C-helix (88-99)

Stable and unstable mutant lysozymes in long helices B and C were constructed to evaluate the effect of the helices on amyloid fibril formation at pH 2. Stable mutant N27D and unstable mutant K33D in the B-helix did not change in amyloid fibril formation. In contrast, stable mutant N93D and unstable mutant K97D in the C-helix showed big differences in behavior as to amyloid fibril formation. Stable mutant N93D showed a longer lag phase of aggregation and suppressed the amyloid fibril formation, whereas unstable mutant K97D showed a shorter lag phase of aggregation and accelerated amyloid fibril formation. These results suggest that the long C-helix is involved mainly in the alpha-helix to beta-sheet transition during amyloid formation of lysozyme.     

Mutations in the B1 domain of protein G that delay the onset of amyloid fibril formation in vitro

We previously reported that under certain experimental conditions, many variants of the B1 domain of IgG-binding protein G from Streptococcus form fibrils reproducibly. The variant I6T53 was the focus of the present study because the lag phase in the kinetics of fibril formation by this variant is significantly longer than that of other variants. This lag phase is distinguished by changes in both intrinsic fluorescence intensity and in light scattering of the protein. NMR diffusion measurements suggest that the soluble protein during the lag phase is monomeric. The kinetic profiles of fibril formation are found to depend on experimental conditions. The first kinetic phase diminishes almost completely when the reaction is seeded with preformed amyloid fibrils.     

Kinetics of activation of L-lactate dehydrogenase from Streptococcus lactis by fructose 1,6-bisphosphate

A lag is observed before the steady state during pyruvate reduction catalysed by lactate dehydrogenase from Streptococcus lactis. The lag is abolished by preincubation of enzyme with the activator fructose 1,6-bisphosphate before mixing with the substrates. The rate constants for the lag phase showed a linear dependence on fructose-1,6-bisphosphate concentration, with a second-order rate constant of 2.0 X 10(4) M-1 s-1, but were independent of enzyme concentration. Binding of fructose 1,6-bisphosphate produces a decrease in the protein fluorescence of the enzyme. The second-order rate constant for the fluorescence change is twice that for the lag in pyruvate reduction. The results suggest that binding of fructose 1,6-bisphosphate induces a conformational change in the enzyme, producing a form with reduced protein fluorescence and increased activity towards pyruvate reduction.     

Thermal and acid denaturation of bovine lens α-crystallin

The chaperone-like protein α-crystallin is a ～35 subunit hetero-oligomer consisting of αA and αB subunits in a 3:1 molar ratio and has the function of maintaining eye lens transparency. We studied the thermal denaturation of α-crystallin by differential scanning calorimetry (DSC), circular dichroism (CD), and dynamic light scattering (DLS) as a function of pH. Our results show that between pH 7 and 10 the protein undergoes a reversible thermal transition. However, the thermodynamic parameters obtained by DSC are inconsistent with the complete denaturation of an oligomeric protein of the size of α-crystallin. Accordingly, the CD data suggest the presence of extensive residual secondary structure above the transition temperature. Within the pH range from 4 to 7 the increased aggregation propensity around the isoelectric point (pI ～ 6) precludes observation of a thermal transition. As pH decreases below 4 the protein undergoes a substantial unfolding. The secondary structure content of the acid-denatured state shows little sensitivity to heating. We propose that the thermal transition above pH 7 and the acid-induced transition at ambient temperature result in predominant denaturation of the αB subunit. Although the extent of denaturation of the αA subunit cannot be estimated from the current data, the existence of a native-like conformation is suggested by the preserved association of the subunits and the chaperone-like activity. A key difference between the thermal and the acid denaturation is that the latter is accompanied by dissociation of αB subunits from the remaining αA-oligomer, as supported by DLS studies.     

CHOLERA TOXIN

1. Death in several infectious diseases is caused by protein toxins secreted by invading bacteria. Cholera toxin is a simple protein secreted by Vibrio cholerae colonizing the gut; it is responsible for the massive diarrhoea that is cholera. 2. The primary action of cholera toxin is an activation of adenylate cyclase, an enzyme found on the inner membrane of eukaryotic cells that catalyses the conversion of ATP to cyclic AMP. Consequent increases in the intracellular concentration of cyclic AMP are responsible for other manifestations of cholera toxin including the diarrhoea. The toxin is active on almost all eukaryotic cells. 3. The toxin can be purified from culture filtrates of V. cholera. It has a molecular weight of 82000; and is composed of one subunit A (itself two polypeptide chains joined by a disulphide bond: AI (22000) and A2 (5000)) and five subunits B (11500). These can be separated in dissociating solvents such as detergents or 6 M guanidine hydrochloride. An amino-acid sequence of subunit B has been published. The five B subunits (sometimes found by themselves in the filtrate and known as ‘choleragenoid’) are probably arranged in a ring with the subunit A in the middle joined to them non-covalently by peptide A2. 4. The first action of cholera toxin on a cell is to bind to the membrane strongly and irreversibly. Several thousand molecules of toxin bind to each cell and the binding constants are of the order of 10^-10 M. The binding is rapid, but is followed by a lag phase of about an hour before the intracellular cyclic AMP concentration begins to increase. 5. Ganglioside G_M1, a complex amphiphilic lipid found in cell membranes, binds tightly to the toxin which shows an enzyme-like specificity for this particular ganglioside. Toxin that has already bound ganglioside can no longer bind to cells and is therefore inactive. This and other experiments using cells depleted of endogenous ganglioside suggest that ganglioside G_M1 is the natural receptor of the toxin on the cell surface. The binding is followed by a lateral movement of the toxin-ganglioside complex in the cell surface forming a ‘cap’ at one pole of the cell. 6. The binding of ganglioside by toxin is a function exclusively of subunit B; Subunit A does not bind and can be eluted with 8 M urea from an insolubilized toxin-ganglioside complex. Subunit B is not by itself active, and so preincubation with B can protect cells or even whole gut from the action of toxin by occupying all the ganglioside binding sites. 7. Subunit A is responsible for activation of adenylate cyclase. Purified subunit A or just peptide AI is active by itself and this activity is not inhibited by ganglioside or by antisera to subunit B. In intact cells the activity is low and shows the characteristic lag phase but in lysed cells the subunit (or the whole toxin) is much more active and there is no lag phase. This suggests that the lag phase represents the time that subunit A takes to cross the cell membrane and get to its target. 8. Several cofactors are needed for toxin activity in lysed cells: NAD+, ATP, sulphydryl compounds and another unidentified cytoplasmic component. The activity of the cyclase is altered in a complex way generally rather similarly to the action of hormones such as adrenalin, but it is difficult to draw any general conclusions. 9. There are two chief theories of how cholera toxin acts. The first is that subunit A (or just peptide AI) enters the cell and there catalyses some reaction leading to activation of the cyclase. The cleavage of NAD+ into nicotinamide and adenosine diphosphoribose could be such a reaction; it is catalysed by high concentrations of cholera toxin. 10. The other theory is that part of the toxin binds directly to the adenylate cyclase or to some other molecule that can then interact with the cyclase, perhaps after the lateral movement of the toxin-ganglioside complex in the cell surface. This binding may be related to the known action of guanyl nucleotides on the cell surface. 11. The entry of peptide AI into the cell and its transport through the membrane is mediated by the binding of subunits B to the cell surface, perhaps just because the binding increases the local concentration of subunit A, or perhaps following specific conformational changes in the subunits and the formation of a tunnel of B subunits through the membrane. An experiment showing that the toxin remains active when the subunits are covalently bonded together suggests that peptide AI does not separate completely from the rest of the molecule. 12. There are several other proteins that resemble cholera toxin in structure and function. For example, glycoprotein hormones such as thyrotrophin also activate adenylate cyclase and have an apparently similar subunit structure with one type of subunit that binds to a ganglioside. There may also be analogies between the amino-acid sequences of toxin and hormones. 13. The enterotoxin made by some strains of Escherichia coli produces a similar diarrhoea to that of cholera. Several different toxic proteins have been prepared but they all seem to activate adenylate cyclase in the same sort of way as cholera toxin does and also to cross-react immunologically with it. The E. coli toxin also reacts with ganglioside G, but the reaction is weak and probably physiologically insignificant. Salmonella typhimurium secretes a similar toxin. 14. Tetanus toxin also reacts with a ganglioside receptor. This protein has two polypeptide chains of which only one reacts with the ganglioside; but the molecular activity is not yet known. 15. Diphtheria toxin has an A fragment that is directly responsible for the toxicity (by catalysing an NAD+ cleavage reaction leading to the total inhibition of protein synthesis) and a B fragment that gets the A fragment into the cells. This structure of active and binding components therefore seems to be common to many toxins. 16. The ability to produce toxin may confer some selective advantage on V. cholerae. The toxin may originate from accidental incorporation of DNA from an eukaryotic host, or alternatively from some material involved with the cyclic AMP metabolism of the bacterium.     

BRICHOS Domains Efficiently Delay Fibrillation of Amyloid β-Peptide

Amyloid diseases such as Alzheimer, Parkinson, and prion diseases are associated with a specific form of protein misfolding and aggregation into oligomers and fibrils rich in β-sheet structure. The BRICHOS domain consisting of ～100 residues is found in membrane proteins associated with degenerative and proliferative disease, including lung fibrosis (surfactant protein C precursor; pro-SP-C) and familial dementia (Bri2). We find that recombinant BRICHOS domains from Bri2 and pro-SP-C prevent fibril formation of amyloid β-peptides (Aβ(40) and Aβ(42)) far below the stoichiometric ratio. Kinetic experiments show that a main effect of BRICHOS is to prolong the lag time in a concentration-dependent, quantitative, and reproducible manner. An ongoing aggregation process is retarded if BRICHOS is added at any time during the lag phase, but it is too late to interfere at the end of the process. Results from circular dichroism and NMR spectroscopy, as well as analytical size exclusion chromatography, imply that Aβ is maintained as an unstructured monomer during the extended lag phase in the presence of BRICHOS. Electron microscopy shows that although the process is delayed, typical amyloid fibrils are eventually formed also when BRICHOS is present. Structural BRICHOS models display a conserved array of tyrosine rings on a five-stranded β-sheet, with inter-hydroxyl distances suited for hydrogen-bonding peptides in an extended β-conformation. Our data imply that the inhibitory mechanism is reliant on BRICHOS interfering with molecular events during the lag phase.