This biochemical purification method with mass spectrometry-based proteomic analysis facilitates the robust characterization of amyloid fibril cores, which may accelerate the identification of targets for preventing Alzheimer’s disease.
Proteinaceous fibrillar inclusions are key pathological hallmarks of multiple neurodegenerative diseases. In the early stages of Alzheimer's disease (AD), amyloid-beta peptides form protofibrils in the extracellular space, which act as seeds that gradually grow and mature into large amyloid plaques. Despite this basic understanding, current knowledge of the amyloid fibril structure, composition, and deposition patterns in the brain is limited. One major barrier has been the inability to isolate highly purified amyloid fibrils from brain extracts. Affinity purification and laser capture microdissection-based approaches have been previously used to isolate amyloids but are limited by the small quantity of material that can be recovered. This novel, robust protocol describes the biochemical purification of amyloid plaque cores using sodium dodecyl sulfate (SDS) solubilization with sucrose density gradient ultracentrifugation and ultrasonication and yields highly pure fibrils from AD patients and AD model brain tissues. Mass spectrometry (MS)-based bottom-up proteomic analysis of the purified material represents a robust strategy to identify nearly all the primary protein components of amyloid fibrils. Previous proteomic studies of proteins in the amyloid coronae have revealed an unexpectedly large and functionally diverse collection of proteins. Notably, after refining the purification strategy, the number of co-purifying proteins was reduced by more than 10-fold, indicating the high purity of the isolated SDS insoluble material. Negative staining and immuno-gold electron microscopy allowed confirmation of the purity of these preparations. Further studies are required to understand the spatial and biological attributes that contribute to the deposition of these proteins into amyloid inclusions. Taken together, this analytical strategy is well-positioned to increase the understanding of amyloid biology.
Amyloid is an extremely stable supramolecular arrangement that is found in a diverse panel of proteins, some of which lead to pathological changes1. The accumulation of intra- or extracellular amyloid aggregates is observed in several neurodegenerative diseases2. Amyloid aggregates are heterogeneous and are enriched with a large number of proteins and lipids3. In recent years, interest in the amyloid proteome has generated substantial interest among basic and translational neuroscientists. Several methods have been developed to extract and purify amyloid aggregates from mouse and post-mortem human brain tissues. Laser-capture microdissection, immunoprecipitation, decellularization, and biochemical isolation of amyloid aggregates are widely used methods to extract and purify amyloid plaques, fibrils, and oligomers4,5,6,7. Many of these studies have focused on determining the protein composition of these tightly packed fibrillar deposits using semi-quantitative MS. However, the available results are inconsistent, and the surprisingly large number of co-purifying proteins previously reported are challenging to interpret.
The primary limitation of the existing literature describing the amyloid core proteome in AD and AD mouse model brains is that the purified material contains an unmanageable number of co-purifying proteins. The overall goal of this method is to overcome this limitation and develop a robust biochemical purification for isolating amyloid fibril cores. This strategy employs a previously described sucrose density gradient ultracentrifugation-based biochemical method for the isolation of SDS insoluble enriched amyloid fractions from post-mortem AD human and mouse brain tissues8,9. This method builds on the existing literature but goes further with ultrasonication and SDS washes to remove most of the loosely bound amyloid-associated proteins, leading to the isolation of highly purified amyloid fibrils (Figure 1). The fibrils purified by this protocol overcome several existing challenges frequently encountered in structural studies of amyloid fibrils isolated from brain extracts. Visualization of these fibrils with transmission electron microscopy (TEM) confirms the integrity and purity of the purified material (Figure 2). In this study, the isolated fibrils are solubilized and digested to peptides with trypsin, and label-free MS analysis can readily reveal the identity of the proteins forming the fibril core. Notably, some of these proteins have an inherent tendency to form supramolecular assemblies in non-membrane-bound organelles. In addition, many of the proteins identified in the analysis of amyloid-beta (Aβ) fibrils are also associated with other neurodegenerative diseases, suggesting that these proteins may play a key role in multiple proteinopathies.
This SDS/ultrasonication method is unlikely to alter or disrupt the structure of the fibril cores. The purified material is also suitable for a wide range of top-down and bottom-up proteomic analysis approaches and additional MS-based structural analysis strategies, such as chemical crosslinking or hydrogen-deuterium exchange. The overall recovery using this method is relatively high and, thus, is suitable for detailed structural studies, which require micrograms to milligrams of the purified material. The purified material is also suitable for structural studies using cryoEM and atomic force microscopy. This protocol, in combination with the stable isotopic labeling of mammals, can facilitate solid-state nuclear magnetic resonance (NMR) studies of amyloid structure10.
This protocol involves the use of human or vertebrate brain tissues. All the research was performed in compliance with the approved institutional guidelines of the Northwestern University. The current workflow is standardized using APP-knock in (AppNL-G-F/NL-G-F) mouse brain cortical and hippocampal brain region extracts11. This protocol has been optimized for brain extracts from mice at 6-9 months of age, and it can effectively purify amyloids from both male and female animals.
NOTE: For a better understanding of the overall experimental procedure, see Figure 1 for a schematic of the workflow.
1. Tissue harvesting and amyloid purification
NOTE: Ideally, amyloid fibrils should be isolated from freshly dissected brain regions. However, this method also works well with snap- or flash-frozen brain tissues. Below is a brief outline of snap-freezing brain tissues for storage for use at a later time.
2. Enrichment of SDS insoluble material
NOTE: Perform all the steps on ice and centrifuge at 4 °C, unless stated otherwise. Details of all the buffers and solutions used in this protocol are provided in Supplementary File 1. Manufacturers and catalog numbers of chemicals and instruments are provided in the Table of Materials.
3. Amyloid purification
NOTE: Combine the two pellets, solubilize by pipetting until obtaining a uniform solution and proceed with the following steps of amyloid purification.
4. Methanol chloroform precipitation
NOTE: If the final goal is to perform protein analysis, it is recommended to desalt and remove additional non-proteinaceous impurities.
5. Trypsin digestion
6. Peptide cleanup
7. Setting up mass spectrometer for peptide analysis
NOTE: For MS parameters, see Supplementary File 1 (adapted from a previous publication from the lab)14.
8. MS data analysis
Here, a detailed method for the isolation and purification of amyloid fibrils using a modified sucrose density gradient ultracentrifugation purification method is summarized (see Figure 1). The innovation in this method is the inclusion of steps of ultrasonication-based washing using a water bath sonication system followed by SDS solubilization, which removes many loosely associated proteins from the amyloid fibrils that co-purify with the highly dense and clean fibrils. The ultrasonication step generates a high shearing force and agitates the fibrils, loosening the hydrophobic forces and releasing the SDS soluble loosely associated proteins into the SDS wash buffer. In turn, small quantities of highly pure amyloid fibril cores are recovered. As shown in Figure 2A, a visible pellet, which initially appears opaque (possibly due to impurities), can be seen after enrichment; however, following the ultrasonication and multiple SDS washes, the pellet turns semi-transparent and is hardly visible. The representative Congo red staining of purified amyloids as compared to the SDS soluble fraction documents the enrichment of the amyloid fibrils (Figure 2B). Congo red staining can be used to confirm the amyloid material in different fractions and can be visualized using a bright field microscope. As shown in Figure 2C, the SDS soluble fraction does not stain with the Congo red dye.
The structure of the purified fibrils with negative staining transmission electron microscopy analysis confirmed the presence of nearly pure amyloid fibrils (Figure 2D). Additionally, immunogold labeling using a combination of Aβ42 (6E10 and 4G8) antibodies confirmed the presence of Aβ42 peptides (Figure 2E). To investigate the composition and structural features of the purified material, we used immunoblot techniques for Aβ peptides and hallmark structural signatures (e.g., fibrils). Representative dot blot analysis of the fractions collected during amyloid isolation showed a relative abundance of Aβ42 peptides and fibrils using anti-Aβ42 and anti-fibril (LOC) antibodies (Figure 3A). Similarly, the western blot of the representative fractions also showed enrichment of Aβ42-containing fibrils in high molecular weight proteins trapped in the wells of the SDS PAGE gel (Figure 3B). To understand the composition of these high molecular weight fibrils, purified amyloid fractions were subjected to MS-based proteomic analysis. These semi-quantitative results revealed the presence of approximately 250 proteins, while the fraction collected before ultrasonication and SDS washes contained more than 2500 proteins (Figure 3C). This indicates the effectiveness of these two crucial steps that are included in this purification protocol. Taken together, multiple independent results indicate the high abundance of similar protein classes in fibril cores.
Gene Ontology (GO) cellular component analysis for one representative MS dataset in Figure 4 revealed that a large number of proteins present in the fibril cores are associated with non-membrane-bound organelle and supramolecular complexes. This observation is likely due to the inherent tendency of many proteins to aggregate themselves or co-aggregate with other proteins in proteinaceous inclusions that are in close proximity. Physical forces play crucial roles in these interactions. The other cellular organelle and components primarily represented by these proteins are mitochondria, cytoskeleton, cell membrane, and myelin sheath. Many of these proteins interact with Aβ peptides, oligomers, or protofibrils at different stages of amyloid formation. They might interact close to the plasma membrane, where Aβ peptides are released. The interaction may also happen while some of these proteins are released directly or via vesicular transport into the extracellular space. Secretion or exocytosis of protein aggregates is one among many strategies that cells use to cope and reduce the burden associated with protein aggregates15. This leaves another opportunity where some of the intracellular proteins can bind to Aβ peptides. The protocol provides a framework for further optimization depending on the experimental goals. For example, the purity and yield, by altering the number of ultrasonication and SDS washes, can be fine-tuned accordingly.
Figure 1: A diagrammatic overview of the workflow for isolation of amyloid fibrils core from AD post-mortem human or model animal brain tissues. Please click here to view a larger version of this figure.
Figure 2: Confirmation of amyloid extraction using biochemical staining and imaging of amyloid fibrils. (A) Enriched amyloid-containing SDS insoluble pellet appears opaque off-white in color. (B) Congo red staining of SDS soluble supernatant and SDS insoluble pellet containing purified amyloid blotted on 0.45 µm nitrocellulose membrane; BCA readings were used for normalization of the loading amount of the proteins. BCA assay was performed as per the manufacturer's instructions. (C) Bright-field images of SDS soluble and amyloid material following Congo red staining (scale bar: 100 µm). (D) Visualization of SDS soluble fraction and purified amyloid fibrils using negative staining under the electron microscope (scale bar: 100 nm). (E) Confirmation of Aβ42 peptides abundance in purified amyloid fibrils by immunogold electron microscopy using Aβ42 (6E10 and 4G8) antibodies (scale bar: 50 nm). Please click here to view a larger version of this figure.
Figure 3: Validation of amyloid purification using immunoblot and MS analysis. (A) Dot blot and (B) Western blot analysis of several representative fractions collected during the amyloid purification process using anti-fibril LOC and anti-Aβ42 antibody; BCA assay readings were used for normalization of the loading amount of the proteins. (C) Number of proteins recovered in label-free mass spectrometry analysis of enriched and purified amyloid fractions; micro BCA assay was used for loading 3 µg of digested peptides for each MS analysis. BCA and micro BCA assays were performed as per the manufacturer's instructions. Please click here to view a larger version of this figure.
Figure 4: Gene Ontology analysis of proteins abundant in purified amyloid fractions. (A) Cellular components and (B) KEGG pathways. Please click here to view a larger version of this figure.
Supplementary File 1: Buffers and solutions, mass spectrometry parameters and ProLuCID search parameters for identification of peptides. Please click here to download this File.
Supplementary Table 1- A representative list of m/z ratios identified in MS run for amyloid beta peptides of APP knock in mouse models. Please click here to download this Table.
Developing a clear understanding of amyloid structure and composition is challenging for structural biologists and biochemists due to the biological complexities and experimental limitations in extracting purified fibrils from AD brain tissues16,17. Amyloid fibrils are polymorphic at the molecular level, showing a heterogeneous population of varying lengths and complexities18,19. To better understand their biological features and pathological relevance, an exhaustive characterization of the composition of polymorphic amyloid fibrils obtained from post-mortem human and AD mouse brain tissues is required20,21. A large subset of proteins directly interacts with Aβ42, while others may have a tendency to form large fibrillar structures or protein complexes22,23,24. It is more challenging to determine the roles played by the proteins interacting with Aβ42 in amyloid formation, stabilization, and elongation as it relates to AD pathology. In recent years, several proteomic studies have elucidated the differences and similarities in the protein composition of various supramolecular arrangements, membrane-less organelles, inclusion bodies, and protein aggregates25,26,27. In the early stages of AD, multiple cellular proteins (e.g., Aβ42 and microtubule-associated tau protein and apolipoprotein E) misfold and co-aggregate in multiple brain regions28,29. Amyloidogenic proteinaceous formations are one major pathological hallmark of AD and likely contribute to pathogenesis3.
The purification of amyloid fibrils from diseased human brains is a tedious and challenging task and has multiple limitations. One major drawback of the existing methods is the low purity of extracted material, which often restricts their detailed structural analysis using imaging methods, such as cryoEM. Likewise, NMR studies require a few milligrams of purified material, which also needs to be labeled with heavy isotopes (i.e., 15N)30. Post-mortem human brains can meet the first requirement, as the starting material can be increased up to a few grams of human tissues; however, labeling human brain tissues with heavy isotopes is not possible. On the other hand, labeling the AD mouse models with 15N isotopes is possible (although costly) and is now increasingly common31,32. Our lab has used pulse-chase labeling to study the synaptic protein turnover dynamics during disease and aging brains14. We have also used whole-animal heavy isotope labeling to identify long-lived proteins and understand their physiological relevance in elaborate biological structures33,34. However, for the mouse brain, large quantities of starting material are required to obtain a workable quantity of the fibril cores. This method successfully addresses these issues by improving the yield and purity of the extracted amyloid fibrils by modifying the existing biochemical isolation principles. Therefore, this robust protocol for amyloid fibrils extraction from brain tissues can be readily utilized for cryoEM and NMR-based structural studies.
This method utilizes the existing sucrose density gradient-based subcellular fractionation paradigm and removes nonspecific co-purifying proteins in successive steps. After the removal of cell debris, myelin, DNA, and cellular lipids, the amyloid-rich SDS insoluble pellets are isolated. The incorporation of additional steps of multiple ultrasonication coupled SDS washes helps to reduce the numbers of co-purifying cellular components, loosely bound proteins, and smaller SDS soluble polymorphs of amyloids. The final pellet is solubilized in ultrapure water and can be used for many applications, including seeding experiments, biochemical or pharmacological studies, and structural analysis. The purified amyloid fibrils from AD brain tissues are also used to understand the proteomic composition and structural features of fibril cores using MS-based proteomics. This analysis confirmed the presence of a subset of cellular proteins (represented in Figure 4) in the core of the fibrils, which may indicate the possible roles of more than one protein in the formation, elongation, and stabilization of amyloid fibrils over a long course of time. Some of the proteins identified in this analysis are known for their association with more than one neurodegenerative disorder, for example, Adam22, APP, ApoE, β-catenin neurofilament proteins, 14-3-3 proteins, and others.
There are possibilities that some contaminating proteins may appear in proteomic analysis owing to the fact that, following homogenization, some unwarranted interactions may happen among proteins due to the high hydrophobicity of amyloid fibrils. Some of these proteins stick to the cores and are not removed even after multiple rounds of sonication and SDS washing steps. This is one limitation of this amyloid purification strategy. However, it could be addressed using suitable negative controls and performing effective statistical cutoffs in large-scale proteomic studies. Another limitation that we encountered relates to the underestimation of Aβ peptide abundance following trypsin digestion. This has been addressed in this workflow by a targeted MS/MS analysis strategy. GO analysis for KEGG pathways indicates the abundance of proteins belonging to pathways involved in many neurodegenerative diseases, for example, Alzheimer's disease, Parkinson's disease, amyotrophic lateral sclerosis (ALS), and Prion diseases. These proteins are important players in multiple pathological pathways and, thus, have known involvement in disease initiation and progression. Interestingly, some of these proteins require further analysis to determine their possible involvement in the pathology of AD and other neurodegenerative diseases.
Future studies on the pure amyloid material from other disease models may provide an in-depth understanding of the structural patterns and composition of core fibrils and may help in identifying key therapeutic targets.
The authors have nothing to disclose.
This work was supported by the NIH grant R01AG061865 to R.J.V. and J.N.S. The authors thank Vassar and Savas research group members at Northwestern University for their thoughtful discussions. We also sincerely thank Dr(s). Ansgar Seimer and Ralf Langen at the University of South California for their crucial input. We thank Dr. Farida Korabova for sample preparation and negative staining electron microscopy imaging at Northwestern University Center for Advanced Microscopy.
Acclaim PepMap 100 C18 HPLC column 0.075 mm x 20 mm | Thermo Scientific | 164535 | Alternative instruments, chemicals and antibodies from other manufacturers can be used |
Ammonium bicarbonate | Sigma-Aldrich | 9830 | |
anti-amyloid beta (1-16) 6E10 antibody | Biolegend | 803001 | |
anti-amyloid beta (17-24) 4G8 antibody | Biolegend | 800701 | |
anti-amyloid beta (N terminus 82E1) antibody | IBL America | 10323 | |
anti-amyloid fibril LOC antibody | EMD Millipore | AB2287 | |
BCA kit | Thermo Fisher Scientific | 23225 | |
Bioruptor Pico Plus | Diagenode | B01020001 | |
Calcium Chloride | Sigma-Aldrich | C1016 | |
Collagenase | Sigma-Aldrich | C0130 | |
Complete Protease Inhibitor Cocktail | Sigma-Aldrich | 11697498001 | |
Dnase I | Thermo Fisher Scientific | EN0521 | |
EDTA | Sigma-Aldrich | EDS | |
Guanidine hydrochloride | Sigma-Aldrich | G4505 | |
HyperSep C18 Cartridges | Thermo Fisher Scientific | 60108-302 | |
Integrated Proteomics Pipeline – IP2 | http://www.integratedproteomics.com/ | ||
Iodoacetamide (IAA) | Sigma-Aldrich | I1149 | |
K54 Tissue Homogenizing System Motor | Cole Parmer | Glas-Col 099C | |
MaxQuant | https://www.maxquant.org/ | ||
Micro BCA kit | Thermo Fisher Scientific | 23235 | |
Nanoviper 75 μm x 50 cm | Thermo Scientific | 164942 | |
Optima L-90K Ultracentrifuge | Beckman Coulter | BR-8101P-E | |
Orbitrap Fusion TribridMass Spectrometer | Thermo Scientific | IQLAAEGAAPFADBMBCX | |
Pierce C18 Spin Columns | Thermo Fisher Scientific | 89870 | |
Precellys 24 tissue homogenizer | Bertin Instruments | P000062-PEVO0-A | |
ProteaseMAX(TM) Surfactant Trypsin Enhancer | Promega | V2072 | |
RawConverter | http://www.fields.scripps.edu/rawconv/ | ||
Sodium azide | VWR | 97064-646 | |
Sodium dodecyl sulfate (SDS) | Sigma-Aldrich | 74255 | |
Sorvall Legend Micro 21R Microcentrifuge | Thermo Fisher Scientific | 75002446 | |
Speed Vaccum Concentrator | Labconco | 7315021 | |
Tris-2-carboxyethylphosphine (TCEP) | Sigma-Aldrich | C4706-2G | |
Tris-HCl | Thermo Fisher Scientific | 15568025 | |
Trypsin Gold-Mass spec grade | Promega | V5280 | |
UltiMate 3000 RSLCnano System | Thermo Scientific | ULTIM3000RSLCNANO |