Inhibition of Alzheimer beta-peptide fibril formation by serum amyloid P component.

A 39-43-amino acid residue-long fragment (β-peptide) from the amyloid precursor protein is the predominant component of amyloid deposits in the brain of individuals with Alzheimer's disease. Serum amyloid P component (SAP) is present in all types of amyloid, including that of Alzheimer's disease. We have used an in vitro model to study the effects of purified SAP on the fibril formation of synthetic Alzheimer β-peptide 1-42. SAP was found to inhibit fibril formation and to increase the solubility of the peptide in a dose-dependent manner. At a 5:1 molar ratio of Aβ1-42 peptide to SAP, fibril formation was completely inhibited, and approximately 80% of the peptide remained in solution even after 4 days of incubation. At lower SAP concentrations, e.g. at peptide to SAP ratio of 1000:1, short fibrillar like structures, lacking amyloid characteristics, were formed. These structures frequently contained associated SAP molecules, suggesting that SAP binds to the polymerizing peptide in a reaction which prevented further fibril formation.

A 39 -43-amino acid residue-long fragment (␤-peptide) from the amyloid precursor protein is the predominant component of amyloid deposits in the brain of individuals with Alzheimer's disease. Serum amyloid P component (SAP) is present in all types of amyloid, including that of Alzheimer's disease. We have used an in vitro model to study the effects of purified SAP on the fibril formation of synthetic Alzheimer ␤-peptide 1-42. SAP was found to inhibit fibril formation and to increase the solubility of the peptide in a dose-dependent manner. At a 5:1 molar ratio of A␤1-42 peptide to SAP, fibril formation was completely inhibited, and approximately 80% of the peptide remained in solution even after 4 days of incubation. At lower SAP concentrations, e.g. at peptide to SAP ratio of 1000:1, short fibrillar like structures, lacking amyloid characteristics, were formed. These structures frequently contained associated SAP molecules, suggesting that SAP binds to the polymerizing peptide in a reaction which prevented further fibril formation.
Serum amyloid P component (SAP) 1 is a calcium-dependent lectin, the best defined specificity of which is 4,6-cyclic pyruvate acetal of ␤-D-galactose. The SAP pentamer consists of five identical 25-kDa subunits each of 204 amino acids, the threedimensional structure of which was determined recently (1). Each subunit was found to be constructed from multiple antiparallel ␤-strands arranged in two sheets, and the tertiary fold was remarkably similar to that of the legume lectins. It binds to DNA, to chromatin, and to glycosaminoglycans such as heparin, heparan, and dermatan sulfate, which are frequently associated with amyloid deposits (2,3). It is also present in glomerular basal membranes and associated with elastic fibers (4). In many SAP-ligand interactions, phosphorylated and/or sulfated groups are involved (5)(6)(7)(8). After calcium-dependent self-aggregation, or when bound to chromatin, SAP may trigger complement activation (9 -11). In addition, SAP also inhibits the complement regulatory function of C4b-binding protein, a protein to which it is complexed in blood (12,13). Despite the numerous properties ascribed to SAP, its physiological function is largely unknown. SAP is a universal constituent of amyloid deposits, including plaques from Alzheimer's disease (AD), amorphous ␤ deposits, and neurofibrillar tangles (14).
The predominant component of amyloid deposits in the brain of individuals with AD is a 39 -43-amino acid-long peptide (A␤-peptide) which is a proteolytic product of the amyloid precursor protein (15,16). Recently, it was shown that the A␤1-42 and A␤1-43 forms are specifically found in all kinds of AD plaques, indicating that those forms are critically important in AD pathology (17). ␤-Amyloid-related peptides are secreted by cultured cells and are normally present in the cerebrospinal fluid (18). The local fibril formation is probably a multistep process which is influenced by the rate of A␤-peptide production and also involves conformational changes of the peptide. The presence of amyloid-associated proteins such as ␣ 1 -antichymotrypsin (ACT) and apolipoprotein E (apoE) appears to enhance fibril formation (19 -22). ACT is present only in amyloid plaques from AD, while a specific allele of apoE (apoE4) has been associated with late onset AD. It has been suggested that ACT and apoE may act as pathological chaperones promoting fibril formation (22). Although SAP is a universal constituent of all types of amyloid, its physiological role in fibril formation is not known.
SAP binds to preformed amyloid fibrils and to Alzheimer A␤1-40 peptide in vitro in a calcium-dependent reaction (23,24). It has been suggested that SAP protects amyloid from proteolytic degradation in vivo by binding to fibrils and masking fibrillar conformation (1,25). We now wish to report that SAP inhibits amyloid fibril formation from the Alzheimer A␤1-42 peptide in an in vitro model.
Purification of SAP-SAP was prepared from human plasma by barium-citrate absorption followed by chromatography on DEAE-Sephacel and heparin-Sepharose, as described previously (26). The purified protein gave a single band of 25 kDa in SDS-polyacrylamide gel electrophoresis. Three different batches of SAP were tested, and they gave similar results. The concentration of SAP was determined by absorbance at 280 nm using an extinction coefficient of 18.2 (27).
Turbidity Assay-The A␤1-42 (20 l of a 40 M solution) was mixed with different amounts of SAP in 15 mM Tris/HCl, 150 mM NaCl, pH 7.4 (TBS; final volume 500 l) and incubated at 37°C. The final concentration of A␤1-42 was kept constant at 1.6 M, SAP concentrations being 1.6 nM (1000:1), 16 nM (100:1), 32 nM (50:1), and 320 nM (5:1); A␤1-42:SAP ratios are given in parentheses. At different times, the light scattering at 400 nm was measured and compared with controls not containing the peptide. Presented results represent the mean of two experiments.
Electron Microscopy-Samples were applied to carbon-coated copper grids, negatively stained with 2% (v/w) uranyl acetate and examined in a JEOL 200CX electron microscope. The results shown are representative of samples taken from each test solution.
Radioassay-The A␤1-42 peptide was labeled with 125 I using the lactoperoxidase method according to Thorell and Larson (28) and stored in TBS containing 0.02% NaN 3 . The radiolabeled peptide was mixed No influence on A␤1-42 solubility by bovine serum albumin was found. In a control experiment, the radioactivity remaining in solution was determined after a 1-h centrifugation of the samples at 120,000 ϫ g in an Aerofugue TM (Beckman).

RESULTS AND DISCUSSION
To examine the effect of SAP on A␤1-42 fibril formation, an in vitro model was used in which the Alzheimer A␤1-42 peptide spontaneously adopts a␤-pleated sheet conformation and forms elongated, approximately 7-8 nm thick, fibrils. The fibril formation of the A␤1-42 peptide was monitored by light-scattering and electron microscopy (Fig. 1). In the absence of SAP, the light scatter increased to a maximum after 72 h (Fig. 1A), and elongated fibrils, having amyloid characteristics (29), were observed in the electron micrographs (Fig. 1B). The addition of SAP resulted in a dose-dependent inhibition of fibril formation. At a peptide to SAP ratio of 5:1, there was almost no increase in light scattering (Fig. 1A), and the fibril formation was completely inhibited (Fig. 1B). A distinct inhibitory effect was observed also at the highest molar ratios of peptide over SAP. Even at a peptide to SAP ratio of 1000:1, the solubility of the peptide was increased as compared to the control, and short, flexible fibrils were formed as revealed by electron microscopy (Fig. 1D) and as reflected by a slight but significant increase in light scatter (Fig. 1A). These fibrils were 1-2 nm thicker than those formed by A␤1-42 alone. Moreover, they did not exhibit green birefringence after Congo red staining, which is characteristic of amyloid fibrils.
Although the precise mechanism of the molecular interaction between SAP and A␤1-42 has not been elucidated, the changes in fibril morphology suggest that SAP affects the packing mechanism of A␤1-42 and disturbs the typical uniformity of the fibrils. SAP is unique in its ability to inhibit fibril formation at all concentrations reported in this study. The amyloid-associated proteins apoE and ACT have been reported either to inhibit or to stimulate the ␤-peptide amyloid fibril formation. At a low ratio between the A␤1-42 peptide and apoE (1000:1), a significant delay in the onset of amyloid fibril formation was observed, whereas under other conditions, apoE stimulated fibril formation (20,22,30). Similar results have been reported for ACT (19,21).
To study the effect of SAP on the overall solubility of A␤1-42 peptide, the amount of radiolabeled peptide remaining in solution after centrifugation and filtration was measured at various times of incubation (Fig. 2). After a 96-h incubation, the solubility of A␤1-42 peptide reached its minimum with 20% of radioactivity remaining in solution. Addition of increasing amounts of SAP resulted in increased solubility demonstrating that SAP may be able to prevent A␤1-42 aggregation. To ensure that the experimental conditions (10-min centrifugation at 13,000 ϫ g and filtration through 0.2-m filters) were sufficient to remove small aggregates of A␤1-42 peptide, a second experiment was performed in which the peptide:SAP mixtures after a 72-h incubation were centrifuged at 120,000 ϫ g for 1 h. The results confirmed the ability of SAP, at an A␤1-42:SAP ratio of 5:1, to keep essentially all the peptide in solution (89%) for up to 72 h. At A␤1-42:SAP ratios of 100:1 and 1000:1, the amounts of peptide remaining in solution were 81 and 60%, respectively. In this experiment, 54% of the peptide remained Direct binding between SAP and the A␤1-42 peptide was demonstrated by nondenaturing agarose gel electrophoresis, but, in contrast to the complexes between ACT and the A␤1-42 peptide (21), complexes between SAP and the ␤-peptide were not stable in SDS (results not shown). Recently, it has been shown that SAP is able to bind to A␤1-40 immobilized in microtiter plates (24). Together, these data suggest that SAP is able to bind A␤1-42 under nondenaturing conditions.
The inhibitory effect of SAP on Alzheimer A␤1-42 peptide fibril formation appears to reflect a general ability of SAP to inhibit amyloid fibril formation. In support for this concept, SAP was found to inhibit the formation of fibrils from a ␤-pleated sheet containing peptide derived from ␣ 1 -antitrypsin (31). Complete inhibition of fibril formation was observed at a peptide to SAP ratio of 5:1. Even at a peptide to SAP ratio of 1000:1, a clear attenuation of fibril formation was observed (Fig. 3). It was noteworthy that at this latter experimental condition, the morphology of the aggregated peptide was different from that observed in the experiment using A␤1-42 (compare Figs. 1D and 3C). Instead of short flexible fibers, dense aggregates of fibers were observed.
Our present data suggest SAP both to impede the seeding process, which initiates fibril formation (32), and to inhibit fibril growth by binding to the peptide and thus preventing peptide-peptide polymerization. Both processes presumably involve binding of SAP to ␤-pleated sheet structures formed by the polymerizing peptide. If the physiological function of SAP is to inhibit amyloid fibril formation, this process is probably imbalanced in AD, either due to changes in the metabolism of the ␤-peptide or to the presence of other SAP ligands in the plaques, such as glycosaminoglycans which would inhibit SAP function. Moreover, in such pathological situations, the attachment of SAP to amyloid fibrils may lead to increased resistance to proteolysis (1,25). SAP has been extremely well conserved through evolution (33) and no deficiency of SAP has been described, suggesting SAP to have important functions. The ability of SAP to inhibit pathological deposition of amyloid-forming peptides may be such a function.