- Email: [email protected]
Sedimentation Velocity Analysis of the Size Distribution of Amyloid Oligomers and Fibrils
CHAPTER ELEVEN
Sedimentation Velocity Analysis of the Size Distribution of Amyloid Oligomers and Fibrils Yee-Foong Mok, Geoffrey J. Howlett1, Michael D.W. Griffin1 Department of Biochemistry and Molecular Biology, Bio21 Molecular Science and Biotechnology Institute, University of Melbourne, Parkville, Victoria, Australia 1 Corresponding authors: e-mail address: [email protected]; [email protected]
Contents 1. Introduction 2. Preparative Ultracentrifugation 2.1 Separation of Fibrils that Differ in Morphology 2.2 Separation of Linear and Closed-Loop Fibrils 2.3 Isolation of Huntingtin Oligomers and Aggregates 3. Analytical Ultracentrifugation 3.1 Sedimentation Velocity Theory 3.2 Sedimentation Velocity Procedures 4. Heterogeneous Systems 4.1 Effects of αB-Crystallin on Mature ApoC-II Amyloid Fibrils 4.2 The Use of Fluorescence Detection 5. Summary Acknowledgments References
242 243 243 245 245 246 247 250 251 251 252 254 254 254
Abstract Amyloid fibrils result from the self-assembly of proteins into large aggregates with fibrillar morphology and common structural features. These fibrils form the major component of amyloid plaques that are associated with a number of common and debilitating diseases, including Alzheimer's disease. While a range of unrelated proteins and peptides are known to form amyloid fibrils, a common feature is the formation of aggregates of various sizes, including mature fibrils of differing length and/or structural morphology, small oligomeric precursors, and other less well-understood forms such as amorphous aggregates. These various species can possess distinct biochemical, biophysical, and pathological properties. Sedimentation velocity analysis can characterize amyloid fibril formation in exceptional detail, providing a particularly useful method for resolving the complex heterogeneity present in amyloid systems. In this chapter, we describe analytical methods for accurate quantification of both total amyloid fibril formation and the formation of distinct amyloid structures based on differential Methods in Enzymology, Volume 562 ISSN 0076-6879 http://dx.doi.org/10.1016/bs.mie.2015.06.024
#
2015 Elsevier Inc. All rights reserved.
241
242
Yee-Foong Mok et al.
sedimentation properties. We also detail modern analytical ultracentrifugation methods to determine the size distribution of amyloid aggregates. We illustrate examples of the use of these techniques to provide biophysical and structural information on amyloid systems that would otherwise be difficult to obtain.
1. INTRODUCTION Amyloid fibrils form by the self-association of proteins or peptides into large, fibrillar structures with a common cross-β secondary structure core. Over 20 unrelated proteins form amyloid fibrils that are associated with a range of major diseases including Alzheimer’s, and Parkinson’s disease and atherosclerosis (Antoni et al., 2013; Sipe & Cohen, 2000; Zerovnik, 2002). It has become apparent that under different conditions, amyloid proteins aggregate into a multitude of species that vary in size, morphology, and solution properties. For instance, aggregates formed by the β2microglobulin protein are comprised of a heterogeneous mixture of amorphous masses and two types of fibrils with either worm-like or long-straight morphology (Gosal et al., 2005). Studies of how the size and morphologies of β2-microglobulin fibrils change with pH, salt, and temperature conditions provide valuable insights into the competing pathways of amyloid fibril formation. Furthermore, it appears that distinct aggregates formed by specific amyloid-forming proteins affect different cellular processes. Emerging evidence suggests, for example, that small, amyloid oligomers of Aβ are cytotoxic in Alzheimer’s disease neurons, whereas other structural forms of Aβ fibrils are more protective (Petkova et al., 2005). Thus, detailed study of the size distributions of fibrils and aggregates formed by amyloidogenic proteins are vital to understanding the underlying mechanisms of amyloid disease development. Yet, the isolation and biochemical analysis of different amyloid species remain challenging due to their size and relatively insoluble nature. Tackling this problem has led to a range of techniques to characterize the heterogeneity of amyloid fibrils (Bruggink, Muller, Kuiperij, & Verbeek, 2012). Over the last decades, centrifugation technologies have emerged as a robust way to separate and analyze many types of amyloid aggregates. In this chapter, we describe preparative ultracentrifuge methods to isolate specific amyloid aggregates, as well as analytical ultracentrifuge methods for quantifying both total amyloid fibril formation and the size distribution of amyloid fibrils and aggregates. We also illustrate examples of the use of fluorescence detection coupled with analytical ultracentrifugation to
Sedimentation Velocity Analysis of Amyloid Fibrils
243
provide biophysical and structural information on the interaction of nonfibrillar components with amyloid fibrils.
2. PREPARATIVE ULTRACENTRIFUGATION While centrifugation provides the potential to globally examine the size distributions of amyloid aggregates in solution, in some cases, it can be particularly useful to isolate an amyloid species for further study. The size and shape differences between species in a heterogeneous amyloid fibril sample enable the use of defined centrifugal conditions to fractionate various amyloid species in a preparative centrifuge. The theory and factors governing the sedimentation of protein molecules in a centrifugal field have been covered extensively in an earlier review (Mok & Howlett, 2006). In general, the amyloid field has used relatively short periods of low-speed centrifugation around 14,000–16,000 g and aqueous buffers to pellet insoluble fibrils and large aggregates of the major amyloid proteins. Smaller amyloid species (e.g., short fibril fragments or oligomers) can be sedimented using longer periods of high-speed centrifugation (>50,000 g) (Mok & Howlett, 2006). In recent years, we have developed protocols to isolate distinct species within complex amyloid mixtures and present, in this section, three preparative centrifugation regimes used to isolate different sized aggregates formed by apolipoprotein C-II and huntingtin.
2.1 Separation of Fibrils that Differ in Morphology Human apolipoprotein (apo) C-II is an exchangeable apolipoprotein that binds reversibly to the polar lipid surface of plasma lipoprotein particles in vivo and associates with a range of natural and synthetic lipid surfaces in vitro. In the absence of lipid, apoC-II is largely unstructured and readily self-assembles into amyloid fibrils with a characteristic flexible ribbon morphology (Fig. 1A). These fibrils are 11–12 nm in width, micrometers in length, and appear to have recurrent constrictions along their contour length every 30–70 nm (Hatters, MacPhee, Lawrence, Sawyer, & Howlett, 2000). In contrast, at low concentrations of phospholipid micelles, apoC-II aggregates via a two-phase process (Griffin et al., 2008), with aggregate morphology after the second stage resembling straight, inflexible rods (Fig. 1B). These straight fibrils are 13–14 nm wide, with no apparent large-scale constrictions, and appear to be up to several micrometers long (Griffin et al., 2008). Straight fibrils sediment more rapidly than their flexible ribbon counterparts, a property that was exploited to separate straight
244
Yee-Foong Mok et al.
Figure 1 Transmission electron micrographs of apoC-II amyloid fibrils. (A) ApoC-II fibrils alone formed at 1 mg/mL, showing flexible ribbon morphology. (B) ApoC-II fibrils formed in the presence of 500 μM 1,2-di-octanoyl-sn-glycero-3-phosphocholine, showing straight-rod morphology. (C) ApoC-II closed loops purified from preformed fibrils using preparative centrifugation. Panels (A) and (B) adapted with permission from Griffin et al. (2008) and panel (C) adapted with permission from Yang, Griffin, Binger, Schuck, and Howlett (2012).
fibrils from ribbon-type fibrils for further study: fibril samples formed at 1.0 mg/mL (112.1 μM) apoC-II in the presence of 500 μM phospholipid were centrifuged for 2 min at relatively low speeds of 14,500 g and the supernatant removed. The resulting straight fibrils in the pellet were then resuspended in an equal volume of 100 mM phosphate, pH 7.4. The process
Sedimentation Velocity Analysis of Amyloid Fibrils
245
was repeated a total of three times and the isolated straight fibrils in the pellet fraction pooled. Under the centrifugal conditions, the bulk of ribbon-type fibrils remained in the supernatant and was separately pelleted by centrifuging at 350,000 g for 10 min.
2.2 Separation of Linear and Closed-Loop Fibrils Both sedimentation velocity analysis and electron microscope images of ribbon-type apoC-II fibrils have consistently revealed a subpopulation of fibrils that appear as closed-loop structures (Hatters et al., 2000). Similar loop or “ring” structures have been observed in amyloid fibrils formed by apoCIII (de Messieres, Huang, He, & Lee, 2014), κ-casein (Thorn et al., 2005), and α-synuclein (Conway et al., 2000), with the rings in the latter hypothesized to be on-pathway for mature fibril formation. ApoC-II fibril formation proceeds via a reversible model that includes elongation, dissociation, fibril breaking, and fibril rejoining (Binger et al., 2008). At equilibrium, closed loops comprise a small proportion of apoC-II ribbon-type fibrils (Yang et al., 2012) and are thought to form by annealing of the free ends of individual, short fibrils. ApoC-II loops appear similar in width and morphology to flexible ribbons, but have a much shorter average contour length of approximately 292 nm (Yang et al., 2012). To prepare these loop structures for study, it was necessary to first enrich the population of ring structures. ApoC-II ribbon-type fibrils were formed at a protein concentration of 0.3 mg/mL by incubation at 25 °C for 7 days. Mature fibrils (1 mL aliquots) were then freeze–thawed 10 times by submerging fibril samples in liquid nitrogen for 25 s followed by rapid thawing in water at 42 °C. This process serves to fragment large fibrils into shorter lengths that form ring structures more readily. After reannealing at 25 °C for 24 h, the samples were centrifuged at 47,500 g for 8 min using an Optima Max centrifuge and a TL-100.1 rotor (Beckman Coulter, Fullerton, CA) resulting in sedimentation of large linear fibrils and significant enrichment of closed-loop fibrils in the supernatant fraction (Fig. 1C). The centrifugal speed and time used was intermediate between those used for pelleting straight fibrils and those for pelleting flexible ribbon fibrils, and demonstrates the utility of preparative ultracentrifugation for separating distinct amyloid aggregates (Yang et al., 2012).
2.3 Isolation of Huntingtin Oligomers and Aggregates The flexibility of preparative centrifugation is illustrated in a final example of a protocol for huntingtin aggregates. The mutant form of huntingtin exon 1
246
Yee-Foong Mok et al.
protein is associated with the fatal neurological disorder, Huntington’s disease and accumulates as massive inclusion bodies within neurons. Neuronal cell lysates containing mutant huntingtin exon 1 contain heterogeneous populations of inclusions, fibrillar aggregates, monomers, and an invariant pool of oligomers (Olshina et al., 2010; Ormsby, Ramdzan, Mok, Jovanoski, & Hatters, 2013). Isolation of huntingtin oligomers is of interest due to hypotheses that oligomers are the toxic agents of amyloid disease (Slow, Graham, & Hayden, 2006). However, it has been technically difficult to fractionate small populations of oligomers from bulk monomers and inclusions using preparative centrifugation alone. Thus, the following protocol was developed (Ormsby et al., 2013). Neuronal cell lysate containing mutant huntingtin aggregates was centrifuged at 14,000 g for 10 min, which pellets inclusions, oligomers, and cell debris. The supernatant, containing monomers, was decanted and the pellet was resuspended in aqueous buffer. Desalting resin from prepacked 0.8 mL Pierce Zeba centrifuge columns (Thermo Scientific, Waltham, MA) was manually removed and replaced with 300 μL of Sephacryl S1000 gel filtration resin (GE Life Sciences, Pittsburgh, PA). The columns were placed on top of empty microfuge tubes, and the resuspended pellet loaded on to the columns. The columns were centrifuged at 1500 g for 1 min, whereby oligomers passed through the resin and collected in the microfuge tubes, while inclusions and large aggregates remained trapped at the top of the columns. The procedure was repeated three times, and the accumulated oligomers pooled. Here, creative refinement of centrifugal procedures enabled the isolation of a smaller entity within a highly complex mixture of aggregates and cell debris.
3. ANALYTICAL ULTRACENTRIFUGATION The analytical ultracentrifuge has been widely used in sedimentation equilibrium and velocity studies to characterize the size and interactions of macromolecules in solution. Sedimentation equilibrium methods offer the advantage that the system at equilibrium yields molecular weight information directly. However, the very large size of amyloid fibrils precludes this approach, since it is generally not possible to centrifuge at sufficiently low speeds to obtain equilibrium distributions. Nevertheless, sedimentation equilibrium is useful in defining the molecular weights of subunits that form amyloid fibrils (Hammarstrom, Jiang, Deechongkit, & Kelly, 2001; Lashuel et al., 2002; Lashuel, Lai, & Kelly, 1998). On the other hand, sedimentation velocity analysis, which measures the rate of movement of fibrils in a
Sedimentation Velocity Analysis of Amyloid Fibrils
247
centrifugal field, offers a very powerful method to characterize the size and heterogeneity of amyloid fibrils. Typical sedimentation velocity experiments on amyloid fibrils take 3–5 h to complete, permitting the time course of slow amyloid formation to be determined (Binger et al., 2008). Recent studies include analysis of the size distributions of Aβ oligomers (NagelSteger et al., 2010), ultrasonication effects on β2-microglobulin (Chatani et al., 2009), and the analysis of huntingtin aggregates (Ormsby et al., 2013). We have previously reviewed some of the basic information pertaining to the analysis of sedimentation velocity data for amyloid fibrils (Mok & Howlett, 2006; Pham, Mok, & Howlett, 2009). This review established that, at the lowest speed achievable in the analytical ultracentrifuge (about 1500 rpm or 164 g), fibrils with sedimentation coefficients as large as 10,000 can be characterized. Note, however, that some fibrillar aggregates become colloidal, sedimenting rapidly even in very low centrifugal fields, and therefore prohibiting sedimentation velocity analysis.
3.1 Sedimentation Velocity Theory The evolution of the concentration distribution, c(r,t), of a single species of diffusing particles in a spinning rotor is given by the Lamm equation (Fujita, 1962): dc 1 d dc 2 2 (1) ¼ rD sω r c dt r dr dr where r is the radial position, t is the time, c is the concentration, s is the sedimentation coefficient, ω is the angular velocity, and D is the translational diffusion coefficient. The development of rapid desktop computers allowed the use of numeric solutions to the Lamm equation as fitting functions, to fit experimental data to models comprising sedimentation coefficient distributions (c(s)) or several noninteracting components (Demeler & Saber, 1998; Schuck, 1998, 2000). Equation (1) forms the basis for the analysis of the sedimentation velocity of amyloid fibrils. One approach that simplifies data analysis is based on the assertion that the large size of amyloid fibrils reduces their rate of diffusion. Using the assumption that the diffusion coefficient of the fibrils is negligible, the evolution of the concentration distribution of identical nondiffusing particles with a uniform initial distribution is described by a step function: 2 0 f orr < rm eω st 2ω2 st (2) c ðr, t Þ ¼ e 1 else
248
Yee-Foong Mok et al.
where rm is the position of the meniscus. A distribution of nondiffusing and noninteracting particles can be analyzed using an approach referred to as ls-g*(s), where experimental concentration distributions are modeled in a least-squares fashion by a summation of step functions weighted according to a variable distribution of sedimentation coefficients (Schuck, 2000). A computer program called SEDFIT for analyzing sedimentation data includes an ls-g*(s) option and is freely available from http://www. analyticalultracentrifugation.com. For many amyloid fibril systems, electron micrograph images indicate either worm-like or straight rod-like structures (Fig. 1). In the case of worm-like fibrils, explicit equations are available to relate the size of fibrils to their respective diffusion coefficients. The results in Fig. 2 show the dependences of the sedimentation coefficient and diffusion coefficient for apoC-II amyloid fibrils based on their mass per unit length and subunit molecular weights (MacRaild et al., 2003). The results show a strong dependence of the diffusion coefficient for fibrils in the range of 1–2000 subunits reaching a plateau level for larger fibrils. The ability to relate sedimentation coefficients to the molecular weights and diffusion coefficients of fibrils
Figure 2 Dependence of sedimentation coefficient and diffusion coefficient on the number of subunits within a worm-like chain. Parameters used for the worm-like chain are based on the values pertaining to human apoC-II amyloid fibrils.
Sedimentation Velocity Analysis of Amyloid Fibrils
249
permits the use of the program SEDFIT and c(s) analysis to determine the size distribution of diffusing amyloid fibrils (Schuck, 1998, 2000). An example of the use of c(s) for the analysis of apoC-II fibrils is presented in Fig. 3, which shows the fits obtained to experimental data and the resulting size
Figure 3 Sedimentation velocity analysis of apoC-II fibrils formed at different apoC-II concentrations. ApoC-II samples were prepared by incubating apoC-II in refolding buffer (100 mM sodium phosphate, 0.05% sodium azide, pH 7.4) at 25 °C for 14 days prior analytical ultracentrifugation at 8000 rpm at 20 °C with radial scans taken at 8 min intervals. (A) Radial scans for apoC-II sample (0.3 mg/mL), monitored at 280 nm (open circles). The fit of the data to a c(s) model (Schuck, 2000) is also shown (solid lines). (B) Sedimentation coefficient distributions obtained by c(s) analysis for fibril samples formed at 0.1 mg/mL (dotted line), 0.3 mg/mL (solid line), and 0.5 mg/mL (dashed line). Adapted with permission from Yang et al. (2012).
250
Yee-Foong Mok et al.
distributions for fibrils formed at different starting concentrations. These data formed the basis for the development of an isodesmic self-association model coupled to closed-loop formation to describe the formation of apoC-II fibrils (Yang et al., 2012).
3.2 Sedimentation Velocity Procedures Procedures for the analysis of sedimentation data for amyloid fibrils have previously been reviewed (Mok & Howlett, 2006; Pham, Mok, & Howlett, 2009). Briefly, there are a number of considerations. It is important to ensure, when comparing data for amyloid fibrils, that sedimentation experiments are performed under the same conditions of temperature, buffer density, and viscosity. This can be particularly relevant for amyloid fibrils, which often have different structural and solution properties at different temperatures. Experimentally, rotors should be preequilibrated with ultracentrifuge cells assembled ahead of time. The rotor and cells can be subjected to up to 2 h of temperature equilibration in the ultracentrifuge prior to analysis. A recent benchmarking and validation exercise has confirmed the stability of the temperature within the cells when the ultracentrifuge chamber temperature is at 20 °C (Zhao et al., 2014). A further consideration is that large aggregates, such as amyloid fibrils and inclusions, may sediment significantly at 3000 rpm. Thus, minimizing the setup period where the rotor is spinning but data not yet being collected is particularly important. Sedimentation data can be collected using either optical density, interference, or the fluorescence detection system (FDS). For optical density measurements, a wavelength of 280 nm is commonly used although increased sensitivity can be achieved at a wavelength of 230 nm. At this wavelength, advantage is taken of the high absorbance of peptide bonds and high intensity of the light source. Measurements can also be made using the interference optics available on the model XL-I. Interference optics offer the advantage that scans can be collected quickly (<10 s) with high accuracy and sensitivity and are useful for solution conditions where there is interference from strongly absorbing ligands or buffer components. The recent development by AVIV Biomedical of the FDS for the Beckman XL-A analytical ultracentrifuge has greatly increased the versatility and sensitivity of AUC in a broad range of contexts. In particular, the FDS increases the accessible range of analyte concentrations, allowing accurate detection of sedimenting molecules in the picomolar concentration range, and provides specific detection of fluorescent molecules within complex mixtures (Kroe & Laue, 2009). Using this detection system, initiation, laser lock, and setup of the emission gain and experiment settings require the rotor to be spinning
Sedimentation Velocity Analysis of Amyloid Fibrils
251
at 3000 rpm, processes that can take up to 10–15 min to complete. In order to minimize the time of this setup period where fibrils may already start to sediment, fluorescence gains may be set using prior knowledge of the approximate fluorescence yields of samples. Previously saved experimental method conditions may also be used. In addition, firing of the laser requires a centrifuge chamber vacuum below 150 μm. It is thus recommended that the vacuum be allowed to equilibrate to below this value while the rotor is stationary.
4. HETEROGENEOUS SYSTEMS 4.1 Effects of αB-Crystallin on Mature ApoC-II Amyloid Fibrils Continuous sedimentation coefficient (c(s)) distribution analysis can be extended to determine the effects of extrinsic molecules and proteins on the interactions between fibrils. We studied the interaction of the small heat-shock protein, αB-crystallin, with mature apoC-II amyloid fibrils and showed that αB-crystallin binds directly to apoC-II fibrils with an apparent Kd of 5.4 μM (Binger et al., 2013). αB-crystallin exists in solution as a polydisperse and dynamic population of oligomeric complexes, including high-molecular-weight oligomers consisting of more than 50 subunits (Baldwin, Lioe, Robinson, Kay, & Benesch, 2011). Thus, we used analytical ultracentrifugation to investigate the possibility that αB-crystallin oligomers can bind multiple apoC-II fibrils simultaneously thereby mediating their lateral association. Mature apoC-II fibrils formed at an apoC-II concentration of 33 μM were incubated in the presence of increasing concentrations of αBcrystallin up to 15 μM for 72 h and subjected to sedimentation velocity analysis. Sedimentation velocity data (Fig. 4) revealed a significant increase in fibril sedimentation rate in the presence of αB-crystallin at a concentration of 5 μM or greater, indicated by the increased distance between radial scans in these samples. Analysis of the sedimentation velocity data using the c(s) model and conversion of the resulting distributions to weight-average sedimentation coefficients (Sav) confirmed this large increase in sedimentation rate. The calculated weight-average sedimentation coefficient in the presence of 15 μM αB-crystallin was 14-fold that of apoC-II fibrils alone, suggesting that the observed changes in sedimentation behavior were not simply due to binding of αB-crystallin to individual fibrils. Transmission electron microscopy confirmed the presence of large aggregates consisting of many fibrils that were not seen in samples of apoC-II alone, indicating that αB-crystallin induces the lateral association and tangling of apoC-II fibrils.
252
Yee-Foong Mok et al.
Figure 4 Incubation of preformed fibrils with different concentrations of αB-crystallin. Fibrils were incubated with increasing concentrations of αB-crystallin for 72 h and examined by sedimentation velocity and TEM. (A) Radial absorbance scans at 285 nm (circles) of preformed apoC-II fibrils incubated with 0, 0.1, 1, 2, 5, 10, or 15 μM αB-crystallin for 72 h. Fit of the data to the c(s) model is also shown (solid lines). Inset: Sav for each sample. (B) TEM of preformed apoC-II fibrils incubated with 15 μM αBcrystallin for 72 h. Also shown are control samples of apoC-II fibrils incubated for the same period without αB-crystallin (smaller panels). Scale bars correspond to 200 nm. Adapted with permission from Binger et al. (2013).
4.2 The Use of Fluorescence Detection We have recently reviewed the applications of fluorescence detected (FD) AUC to the analysis of amyloid systems, including investigation of amyloidforming proteins labeled with small-molecule fluorophores or green
Sedimentation Velocity Analysis of Amyloid Fibrils
253
Figure 5 Fluorescence-detected sedimentation analysis of apoC-II amyloid fibrils under dilution in the presence and absence of αB-crystallin. (A) Continuous sedimentation coefficient distributions of apoC-II fibrils undiluted or diluted 1:10 and incubated with 0, 2, or 5 μM αB-crystallin for 0 h (solid), 8 h (dotted), 24 h (dashed), or 72 h (dash–dot– dot). Total integrated signal for each distribution was normalized with respect to the corresponding t ¼ 0 sample. Adapted with permission from Binger et al. (2013).
fluorescent protein, and the use of extrinsic fluorophores and fluorescently labeled lipids that bind to amyloid aggregates and intermediates (Mok et al., 2011). We further exploited the ability of FD-AUC to detect the sedimentation of fluorescent components of multicomponent protein mixtures to accurately determine the effect of αB-crystallin on the size distribution of apoC-II fibrils after dilution. Fluorescent apoC-II was prepared by labeling of the protein using Alexa-488 maleimide at an engineered cysteine residue (Ryan, Howlett, & Bailey, 2008), and fibrils were formed from a mixture of 2% Alexa-488-apoC-II and 98% unlabeled, wild-type apoC-II at a total apoC-II concentration of 33 μM. Fluorescent fibrils were then incubated with 0, 2, or 5 μM αB-crystallin and subsequently diluted 1:10 in buffer containing the appropriate concentration of αB-crystallin. After this dilution, the apoC-II concentration was approximately 3.3 μM and, thus, αBcrystallin in the samples comprised a large proportion of the total protein concentration. As high-molecular-weight αB-crystallin oligomers sediment under the centrifugal fields used for analysis of apoC-II fibrils, it was critical to monitor only the fibrillar component of the mixture in order to avoid interference from αB-crystallin and to obtain accurate size distributions and weight-average sedimentation coefficients for the apoC-II fibrils. Analysis of the FD sedimentation velocity data (Fig. 5) showed that dilution of the fibrils in the absence of αB-crystallin resulted in a reduction in the
254
Yee-Foong Mok et al.
weight-average sedimentation coefficient corresponding to fragmentation of fibrils (Binger et al., 2008) and consistent with our model of apoC-II fibril formation including fibril breakage and rejoining (Binger et al., 2008). The presence of αB-crystallin during and after dilution protected fibrils against fragmentation. Interestingly, the sedimentation velocity data also showed a time-dependent reduction in fluorescence signal sedimenting in the presence of αB-crystallin suggesting dissociation of fibrils. Subsequent fibril pelleting assays, performed as described in our earlier review (Mok & Howlett, 2006), confirmed that αB-crystallin mediated the dissociation of fibrils resulting in monomeric apoC-II.
5. SUMMARY The preparative ultracentrifuge provides a useful method for the fractionation of amyloid oligomers, closed loop and linear fibrils, as well as fibrils with different morphologies. Fibril aggregates with sedimentation coefficients up to 10,000 S can be fractionated using this approach. The analytical ultracentrifuge offers much higher precision in the characterization of the size distributions of amyloid fibrils. For fibrils where there is prior knowledge of the relationship between the molecular weight of the fibrils and their diffusion coefficients, c(s) analysis allows the size dependence of fibril diffusion coefficients to be included in the analysis. The recent development of the FDS for the analytical ultracentrifuge has greatly extended the ability to analyze heterogeneous systems and the effects of specific reagents on the sedimentation behavior of amyloid fibrils.
ACKNOWLEDGMENTS We thank Danny Hatters and Angelique Ormsby for permission to include details of the preparation of huntingtin aggregates. This work was supported by the Australian Research Council (projects DP0877800 and DP0984565). M.D.W.G. is the recipient of the C.R. Roper Fellowship and an Australian Research Council Future Fellowship (project number FT140100544).
REFERENCES Antoni, G., Lubberink, M., Estrada, S., Axelsson, J., Carlson, K., Lindsjo, L., et al. (2013). In vivo visualization of amyloid deposits in the heart with 11C-PIB and PET. Journal of Nuclear Medicine: Official Publication, Society of Nuclear Medicine, 54, 213–220. Baldwin, A. J., Lioe, H., Robinson, C. V., Kay, L. E., & Benesch, J. L. (2011). alphaBcrystallin polydispersity is a consequence of unbiased quaternary dynamics. Journal of Molecular Biology, 413, 297–309.
Sedimentation Velocity Analysis of Amyloid Fibrils
255
Binger, K. J., Ecroyd, H., Yang, S., Carver, J. A., Howlett, G. J., & Griffin, M. D. (2013). Avoiding the oligomeric state: AlphaB-crystallin inhibits fragmentation and induces dissociation of apolipoprotein C-II amyloid fibrils. FASEB Journal: Official Publication of the Federation of American Societies for Experimental Biology, 27, 1214–1222. Binger, K. J., Pham, C. L., Wilson, L. M., Bailey, M. F., Lawrence, L. J., Schuck, P., et al. (2008). Apolipoprotein C-II amyloid fibrils assemble via a reversible pathway that includes fibril breaking and rejoining. Journal of Molecular Biology, 376, 1116–1129. Bruggink, K. A., Muller, M., Kuiperij, H. B., & Verbeek, M. M. (2012). Methods for analysis of amyloid-beta aggregates. Journal of Alzheimer’s Disease: JAD, 28, 735–758. Chatani, E., Lee, Y. H., Yagi, H., Yoshimura, Y., Naiki, H., & Goto, Y. (2009). Ultrasonication-dependent production and breakdown lead to minimum-sized amyloid fibrils. Proceedings of the National Academy of Sciences of the United States of America, 106, 11119–11124. Conway, K. A., Lee, S. J., Rochet, J. C., Ding, T. T., Williamson, R. E., & Lansbury, P. T., Jr. (2000). Acceleration of oligomerization, not fibrillization, is a shared property of both alpha-synuclein mutations linked to early-onset Parkinson’s disease: Implications for pathogenesis and therapy. Proceedings of the National Academy of Sciences of the United States of America, 97, 571–576. de Messieres, M., Huang, R. K., He, Y., & Lee, J. C. (2014). Amyloid triangles, squares, and loops of apolipoprotein C-III. Biochemistry, 53, 3261–3263. Demeler, B., & Saber, H. (1998). Determination of molecular parameters by fitting sedimentation data to finite-element solutions of the Lamm equation. Biophysical Journal, 74, 444–454. Fujita, H. (1962). Mathematical theory of sedimentation analysis. New York: Academic Press. Gosal, W. S., Morten, I. J., Hewitt, E. W., Smith, D. A., Thomson, N. H., & Radford, S. E. (2005). Competing pathways determine fibril morphology in the self-assembly of beta2microglobulin into amyloid. Journal of Molecular Biology, 351, 850–864. Griffin, M. D., Mok, M. L., Wilson, L. M., Pham, C. L., Waddington, L. J., Perugini, M. A., et al. (2008). Phospholipid interaction induces molecular-level polymorphism in apolipoprotein C-II amyloid fibrils via alternative assembly pathways. Journal of Molecular Biology, 375, 240–256. Hammarstrom, P., Jiang, X., Deechongkit, S., & Kelly, J. W. (2001). Anion shielding of electrostatic repulsions in transthyretin modulates stability and amyloidosis: Insight into the chaotrope unfolding dichotomy. Biochemistry, 40, 11453–11459. Hatters, D. M., MacPhee, C. E., Lawrence, L. J., Sawyer, W. H., & Howlett, G. J. (2000). Human apolipoprotein C-II forms twisted amyloid ribbons and closed loops. Biochemistry, 39, 8276–8283. Kroe, R. R., & Laue, T. M. (2009). NUTS and BOLTS: applications of fluorescencedetected sedimentation. Analytical Biochemistry, 390, 1–13. Lashuel, H. A., Hartley, D. M., Balakhaneh, D., Aggarwal, A., Teichberg, S., & Callaway, D. J. (2002). New class of inhibitors of amyloid-beta fibril formation. Implications for the mechanism of pathogenesis in Alzheimer’s disease. The Journal of Biological Chemistry, 277, 42881–42890. Lashuel, H. A., Lai, Z., & Kelly, J. W. (1998). Characterization of the transthyretin acid denaturation pathways by analytical ultracentrifugation: Implications for wild-type, V30M, and L55P amyloid fibril formation. Biochemistry, 37, 17851–17864. MacRaild, C. A., Hatters, D. M., Lawrence, L. J., & Howlett, G. J. (2003). Sedimentation velocity analysis of flexible macromolecules: Self-association and tangling of amyloid fibrils. Biophysical Journal, 84, 2562–2569. Mok, Y. F., & Howlett, G. J. (2006). Sedimentation velocity analysis of amyloid oligomers and fibrils. Methods in Enzymology, 413, 199–217.
256
Yee-Foong Mok et al.
Mok, Y. F., Ryan, T. M., Yang, S., Hatters, D. M., Howlett, G. J., & Griffin, M. D. (2011). Sedimentation velocity analysis of amyloid oligomers and fibrils using fluorescence detection. Methods, 54, 67–75. Nagel-Steger, L., Demeler, B., Meyer-Zaika, W., Hochdorffer, K., Schrader, T., & Willbold, D. (2010). Modulation of aggregate size- and shape-distributions of the amyloid-beta peptide by a designed beta-sheet breaker. European Biophysics Journal: EBJ, 39, 415–422. Olshina, M. A., Angley, L. M., Ramdzan, Y. M., Tang, J., Bailey, M. F., Hill, A. F., et al. (2010). Tracking mutant huntingtin aggregation kinetics in cells reveals three major populations that include an invariant oligomer pool. The Journal of Biological Chemistry, 285, 21807–21816. Ormsby, A. R., Ramdzan, Y. M., Mok, Y. F., Jovanoski, K. D., & Hatters, D. M. (2013). A platform to view huntingtin exon 1 aggregation flux in the cell reveals divergent influences from chaperones hsp40 and hsp70. The Journal of Biological Chemistry, 288, 37192–37203. Petkova, A. T., Leapman, R. D., Guo, Z., Yau, W. M., Mattson, M. P., & Tycko, R. (2005). Self-propagating, molecular-level polymorphism in Alzheimer’s beta-amyloid fibrils. Science, 307, 262–265. Pham, C., Mok, Y.-F., & Howlett, G. J. (2009). Sedimentation velocity analysis of amyloid fibrils. Methods in Molecular Biology, 752, 179–196. Ryan, T. M., Howlett, G. J., & Bailey, M. F. (2008). Fluorescence detection of a lipidinduced tetrameric intermediate in amyloid fibril formation by apolipoprotein C-II. The Journal of Biological Chemistry, 283, 35118–35128. Schuck, P. (1998). Sedimentation analysis of noninteracting and self-associating solutes using numerical solutions to the Lamm equation. Biophysical Journal, 75, 1503–1512. Schuck, P. (2000). Size-distribution analysis of macromolecules by sedimentation velocity ultracentrifugation and Lamm equation modeling. Biophysical Journal, 78, 1606–1619. Sipe, J. D., & Cohen, A. S. (2000). Review: History of the amyloid fibril. Journal of Structural Biology, 130, 88–98. Slow, E. J., Graham, R. K., & Hayden, M. R. (2006). To be or not to be toxic: Aggregations in Huntington and Alzheimer disease. Trends in Genetics: TIG, 22, 408–411. Thorn, D. C., Meehan, S., Sunde, M., Rekas, A., Gras, S. L., MacPhee, C. E., et al. (2005). Amyloid fibril formation by bovine milk kappa-casein and its inhibition by the molecular chaperones alphaS- and beta-casein. Biochemistry, 44, 17027–17036. Yang, S., Griffin, M. D., Binger, K. J., Schuck, P., & Howlett, G. J. (2012). An equilibrium model for linear and closed-loop amyloid fibril formation. Journal of Molecular Biology, 421, 364–377. Zerovnik, E. (2002). Amyloid-fibril formation. Proposed mechanisms and relevance to conformational disease. European Journal of Biochemistry/FEBS, 269, 3362–3371. Zhao, H., Balbo, A., Metger, H., Clary, R., Ghirlando, R., & Schuck, P. (2014). Improved measurement of the rotor temperature in analytical ultracentrifugation. Analytical Biochemistry, 451, 69–75.
Sedimentation Velocity Analysis of the Size Distribution of Amyloid Oligomers and Fibrils Yee-Foong Mok, Geoffrey J. Howlett1, Michael D.W. Griffin1 Department of Biochemistry and Molecular Biology, Bio21 Molecular Science and Biotechnology Institute, University of Melbourne, Parkville, Victoria, Australia 1 Corresponding authors: e-mail address: [email protected]; [email protected]
Contents 1. Introduction 2. Preparative Ultracentrifugation 2.1 Separation of Fibrils that Differ in Morphology 2.2 Separation of Linear and Closed-Loop Fibrils 2.3 Isolation of Huntingtin Oligomers and Aggregates 3. Analytical Ultracentrifugation 3.1 Sedimentation Velocity Theory 3.2 Sedimentation Velocity Procedures 4. Heterogeneous Systems 4.1 Effects of αB-Crystallin on Mature ApoC-II Amyloid Fibrils 4.2 The Use of Fluorescence Detection 5. Summary Acknowledgments References
242 243 243 245 245 246 247 250 251 251 252 254 254 254
Abstract Amyloid fibrils result from the self-assembly of proteins into large aggregates with fibrillar morphology and common structural features. These fibrils form the major component of amyloid plaques that are associated with a number of common and debilitating diseases, including Alzheimer's disease. While a range of unrelated proteins and peptides are known to form amyloid fibrils, a common feature is the formation of aggregates of various sizes, including mature fibrils of differing length and/or structural morphology, small oligomeric precursors, and other less well-understood forms such as amorphous aggregates. These various species can possess distinct biochemical, biophysical, and pathological properties. Sedimentation velocity analysis can characterize amyloid fibril formation in exceptional detail, providing a particularly useful method for resolving the complex heterogeneity present in amyloid systems. In this chapter, we describe analytical methods for accurate quantification of both total amyloid fibril formation and the formation of distinct amyloid structures based on differential Methods in Enzymology, Volume 562 ISSN 0076-6879 http://dx.doi.org/10.1016/bs.mie.2015.06.024
#
2015 Elsevier Inc. All rights reserved.
241
242
Yee-Foong Mok et al.
sedimentation properties. We also detail modern analytical ultracentrifugation methods to determine the size distribution of amyloid aggregates. We illustrate examples of the use of these techniques to provide biophysical and structural information on amyloid systems that would otherwise be difficult to obtain.
1. INTRODUCTION Amyloid fibrils form by the self-association of proteins or peptides into large, fibrillar structures with a common cross-β secondary structure core. Over 20 unrelated proteins form amyloid fibrils that are associated with a range of major diseases including Alzheimer’s, and Parkinson’s disease and atherosclerosis (Antoni et al., 2013; Sipe & Cohen, 2000; Zerovnik, 2002). It has become apparent that under different conditions, amyloid proteins aggregate into a multitude of species that vary in size, morphology, and solution properties. For instance, aggregates formed by the β2microglobulin protein are comprised of a heterogeneous mixture of amorphous masses and two types of fibrils with either worm-like or long-straight morphology (Gosal et al., 2005). Studies of how the size and morphologies of β2-microglobulin fibrils change with pH, salt, and temperature conditions provide valuable insights into the competing pathways of amyloid fibril formation. Furthermore, it appears that distinct aggregates formed by specific amyloid-forming proteins affect different cellular processes. Emerging evidence suggests, for example, that small, amyloid oligomers of Aβ are cytotoxic in Alzheimer’s disease neurons, whereas other structural forms of Aβ fibrils are more protective (Petkova et al., 2005). Thus, detailed study of the size distributions of fibrils and aggregates formed by amyloidogenic proteins are vital to understanding the underlying mechanisms of amyloid disease development. Yet, the isolation and biochemical analysis of different amyloid species remain challenging due to their size and relatively insoluble nature. Tackling this problem has led to a range of techniques to characterize the heterogeneity of amyloid fibrils (Bruggink, Muller, Kuiperij, & Verbeek, 2012). Over the last decades, centrifugation technologies have emerged as a robust way to separate and analyze many types of amyloid aggregates. In this chapter, we describe preparative ultracentrifuge methods to isolate specific amyloid aggregates, as well as analytical ultracentrifuge methods for quantifying both total amyloid fibril formation and the size distribution of amyloid fibrils and aggregates. We also illustrate examples of the use of fluorescence detection coupled with analytical ultracentrifugation to
Sedimentation Velocity Analysis of Amyloid Fibrils
243
provide biophysical and structural information on the interaction of nonfibrillar components with amyloid fibrils.
2. PREPARATIVE ULTRACENTRIFUGATION While centrifugation provides the potential to globally examine the size distributions of amyloid aggregates in solution, in some cases, it can be particularly useful to isolate an amyloid species for further study. The size and shape differences between species in a heterogeneous amyloid fibril sample enable the use of defined centrifugal conditions to fractionate various amyloid species in a preparative centrifuge. The theory and factors governing the sedimentation of protein molecules in a centrifugal field have been covered extensively in an earlier review (Mok & Howlett, 2006). In general, the amyloid field has used relatively short periods of low-speed centrifugation around 14,000–16,000 g and aqueous buffers to pellet insoluble fibrils and large aggregates of the major amyloid proteins. Smaller amyloid species (e.g., short fibril fragments or oligomers) can be sedimented using longer periods of high-speed centrifugation (>50,000 g) (Mok & Howlett, 2006). In recent years, we have developed protocols to isolate distinct species within complex amyloid mixtures and present, in this section, three preparative centrifugation regimes used to isolate different sized aggregates formed by apolipoprotein C-II and huntingtin.
2.1 Separation of Fibrils that Differ in Morphology Human apolipoprotein (apo) C-II is an exchangeable apolipoprotein that binds reversibly to the polar lipid surface of plasma lipoprotein particles in vivo and associates with a range of natural and synthetic lipid surfaces in vitro. In the absence of lipid, apoC-II is largely unstructured and readily self-assembles into amyloid fibrils with a characteristic flexible ribbon morphology (Fig. 1A). These fibrils are 11–12 nm in width, micrometers in length, and appear to have recurrent constrictions along their contour length every 30–70 nm (Hatters, MacPhee, Lawrence, Sawyer, & Howlett, 2000). In contrast, at low concentrations of phospholipid micelles, apoC-II aggregates via a two-phase process (Griffin et al., 2008), with aggregate morphology after the second stage resembling straight, inflexible rods (Fig. 1B). These straight fibrils are 13–14 nm wide, with no apparent large-scale constrictions, and appear to be up to several micrometers long (Griffin et al., 2008). Straight fibrils sediment more rapidly than their flexible ribbon counterparts, a property that was exploited to separate straight
244
Yee-Foong Mok et al.
Figure 1 Transmission electron micrographs of apoC-II amyloid fibrils. (A) ApoC-II fibrils alone formed at 1 mg/mL, showing flexible ribbon morphology. (B) ApoC-II fibrils formed in the presence of 500 μM 1,2-di-octanoyl-sn-glycero-3-phosphocholine, showing straight-rod morphology. (C) ApoC-II closed loops purified from preformed fibrils using preparative centrifugation. Panels (A) and (B) adapted with permission from Griffin et al. (2008) and panel (C) adapted with permission from Yang, Griffin, Binger, Schuck, and Howlett (2012).
fibrils from ribbon-type fibrils for further study: fibril samples formed at 1.0 mg/mL (112.1 μM) apoC-II in the presence of 500 μM phospholipid were centrifuged for 2 min at relatively low speeds of 14,500 g and the supernatant removed. The resulting straight fibrils in the pellet were then resuspended in an equal volume of 100 mM phosphate, pH 7.4. The process
Sedimentation Velocity Analysis of Amyloid Fibrils
245
was repeated a total of three times and the isolated straight fibrils in the pellet fraction pooled. Under the centrifugal conditions, the bulk of ribbon-type fibrils remained in the supernatant and was separately pelleted by centrifuging at 350,000 g for 10 min.
2.2 Separation of Linear and Closed-Loop Fibrils Both sedimentation velocity analysis and electron microscope images of ribbon-type apoC-II fibrils have consistently revealed a subpopulation of fibrils that appear as closed-loop structures (Hatters et al., 2000). Similar loop or “ring” structures have been observed in amyloid fibrils formed by apoCIII (de Messieres, Huang, He, & Lee, 2014), κ-casein (Thorn et al., 2005), and α-synuclein (Conway et al., 2000), with the rings in the latter hypothesized to be on-pathway for mature fibril formation. ApoC-II fibril formation proceeds via a reversible model that includes elongation, dissociation, fibril breaking, and fibril rejoining (Binger et al., 2008). At equilibrium, closed loops comprise a small proportion of apoC-II ribbon-type fibrils (Yang et al., 2012) and are thought to form by annealing of the free ends of individual, short fibrils. ApoC-II loops appear similar in width and morphology to flexible ribbons, but have a much shorter average contour length of approximately 292 nm (Yang et al., 2012). To prepare these loop structures for study, it was necessary to first enrich the population of ring structures. ApoC-II ribbon-type fibrils were formed at a protein concentration of 0.3 mg/mL by incubation at 25 °C for 7 days. Mature fibrils (1 mL aliquots) were then freeze–thawed 10 times by submerging fibril samples in liquid nitrogen for 25 s followed by rapid thawing in water at 42 °C. This process serves to fragment large fibrils into shorter lengths that form ring structures more readily. After reannealing at 25 °C for 24 h, the samples were centrifuged at 47,500 g for 8 min using an Optima Max centrifuge and a TL-100.1 rotor (Beckman Coulter, Fullerton, CA) resulting in sedimentation of large linear fibrils and significant enrichment of closed-loop fibrils in the supernatant fraction (Fig. 1C). The centrifugal speed and time used was intermediate between those used for pelleting straight fibrils and those for pelleting flexible ribbon fibrils, and demonstrates the utility of preparative ultracentrifugation for separating distinct amyloid aggregates (Yang et al., 2012).
2.3 Isolation of Huntingtin Oligomers and Aggregates The flexibility of preparative centrifugation is illustrated in a final example of a protocol for huntingtin aggregates. The mutant form of huntingtin exon 1
246
Yee-Foong Mok et al.
protein is associated with the fatal neurological disorder, Huntington’s disease and accumulates as massive inclusion bodies within neurons. Neuronal cell lysates containing mutant huntingtin exon 1 contain heterogeneous populations of inclusions, fibrillar aggregates, monomers, and an invariant pool of oligomers (Olshina et al., 2010; Ormsby, Ramdzan, Mok, Jovanoski, & Hatters, 2013). Isolation of huntingtin oligomers is of interest due to hypotheses that oligomers are the toxic agents of amyloid disease (Slow, Graham, & Hayden, 2006). However, it has been technically difficult to fractionate small populations of oligomers from bulk monomers and inclusions using preparative centrifugation alone. Thus, the following protocol was developed (Ormsby et al., 2013). Neuronal cell lysate containing mutant huntingtin aggregates was centrifuged at 14,000 g for 10 min, which pellets inclusions, oligomers, and cell debris. The supernatant, containing monomers, was decanted and the pellet was resuspended in aqueous buffer. Desalting resin from prepacked 0.8 mL Pierce Zeba centrifuge columns (Thermo Scientific, Waltham, MA) was manually removed and replaced with 300 μL of Sephacryl S1000 gel filtration resin (GE Life Sciences, Pittsburgh, PA). The columns were placed on top of empty microfuge tubes, and the resuspended pellet loaded on to the columns. The columns were centrifuged at 1500 g for 1 min, whereby oligomers passed through the resin and collected in the microfuge tubes, while inclusions and large aggregates remained trapped at the top of the columns. The procedure was repeated three times, and the accumulated oligomers pooled. Here, creative refinement of centrifugal procedures enabled the isolation of a smaller entity within a highly complex mixture of aggregates and cell debris.
3. ANALYTICAL ULTRACENTRIFUGATION The analytical ultracentrifuge has been widely used in sedimentation equilibrium and velocity studies to characterize the size and interactions of macromolecules in solution. Sedimentation equilibrium methods offer the advantage that the system at equilibrium yields molecular weight information directly. However, the very large size of amyloid fibrils precludes this approach, since it is generally not possible to centrifuge at sufficiently low speeds to obtain equilibrium distributions. Nevertheless, sedimentation equilibrium is useful in defining the molecular weights of subunits that form amyloid fibrils (Hammarstrom, Jiang, Deechongkit, & Kelly, 2001; Lashuel et al., 2002; Lashuel, Lai, & Kelly, 1998). On the other hand, sedimentation velocity analysis, which measures the rate of movement of fibrils in a
Sedimentation Velocity Analysis of Amyloid Fibrils
247
centrifugal field, offers a very powerful method to characterize the size and heterogeneity of amyloid fibrils. Typical sedimentation velocity experiments on amyloid fibrils take 3–5 h to complete, permitting the time course of slow amyloid formation to be determined (Binger et al., 2008). Recent studies include analysis of the size distributions of Aβ oligomers (NagelSteger et al., 2010), ultrasonication effects on β2-microglobulin (Chatani et al., 2009), and the analysis of huntingtin aggregates (Ormsby et al., 2013). We have previously reviewed some of the basic information pertaining to the analysis of sedimentation velocity data for amyloid fibrils (Mok & Howlett, 2006; Pham, Mok, & Howlett, 2009). This review established that, at the lowest speed achievable in the analytical ultracentrifuge (about 1500 rpm or 164 g), fibrils with sedimentation coefficients as large as 10,000 can be characterized. Note, however, that some fibrillar aggregates become colloidal, sedimenting rapidly even in very low centrifugal fields, and therefore prohibiting sedimentation velocity analysis.
3.1 Sedimentation Velocity Theory The evolution of the concentration distribution, c(r,t), of a single species of diffusing particles in a spinning rotor is given by the Lamm equation (Fujita, 1962): dc 1 d dc 2 2 (1) ¼ rD sω r c dt r dr dr where r is the radial position, t is the time, c is the concentration, s is the sedimentation coefficient, ω is the angular velocity, and D is the translational diffusion coefficient. The development of rapid desktop computers allowed the use of numeric solutions to the Lamm equation as fitting functions, to fit experimental data to models comprising sedimentation coefficient distributions (c(s)) or several noninteracting components (Demeler & Saber, 1998; Schuck, 1998, 2000). Equation (1) forms the basis for the analysis of the sedimentation velocity of amyloid fibrils. One approach that simplifies data analysis is based on the assertion that the large size of amyloid fibrils reduces their rate of diffusion. Using the assumption that the diffusion coefficient of the fibrils is negligible, the evolution of the concentration distribution of identical nondiffusing particles with a uniform initial distribution is described by a step function: 2 0 f orr < rm eω st 2ω2 st (2) c ðr, t Þ ¼ e 1 else
248
Yee-Foong Mok et al.
where rm is the position of the meniscus. A distribution of nondiffusing and noninteracting particles can be analyzed using an approach referred to as ls-g*(s), where experimental concentration distributions are modeled in a least-squares fashion by a summation of step functions weighted according to a variable distribution of sedimentation coefficients (Schuck, 2000). A computer program called SEDFIT for analyzing sedimentation data includes an ls-g*(s) option and is freely available from http://www. analyticalultracentrifugation.com. For many amyloid fibril systems, electron micrograph images indicate either worm-like or straight rod-like structures (Fig. 1). In the case of worm-like fibrils, explicit equations are available to relate the size of fibrils to their respective diffusion coefficients. The results in Fig. 2 show the dependences of the sedimentation coefficient and diffusion coefficient for apoC-II amyloid fibrils based on their mass per unit length and subunit molecular weights (MacRaild et al., 2003). The results show a strong dependence of the diffusion coefficient for fibrils in the range of 1–2000 subunits reaching a plateau level for larger fibrils. The ability to relate sedimentation coefficients to the molecular weights and diffusion coefficients of fibrils
Figure 2 Dependence of sedimentation coefficient and diffusion coefficient on the number of subunits within a worm-like chain. Parameters used for the worm-like chain are based on the values pertaining to human apoC-II amyloid fibrils.
Sedimentation Velocity Analysis of Amyloid Fibrils
249
permits the use of the program SEDFIT and c(s) analysis to determine the size distribution of diffusing amyloid fibrils (Schuck, 1998, 2000). An example of the use of c(s) for the analysis of apoC-II fibrils is presented in Fig. 3, which shows the fits obtained to experimental data and the resulting size
Figure 3 Sedimentation velocity analysis of apoC-II fibrils formed at different apoC-II concentrations. ApoC-II samples were prepared by incubating apoC-II in refolding buffer (100 mM sodium phosphate, 0.05% sodium azide, pH 7.4) at 25 °C for 14 days prior analytical ultracentrifugation at 8000 rpm at 20 °C with radial scans taken at 8 min intervals. (A) Radial scans for apoC-II sample (0.3 mg/mL), monitored at 280 nm (open circles). The fit of the data to a c(s) model (Schuck, 2000) is also shown (solid lines). (B) Sedimentation coefficient distributions obtained by c(s) analysis for fibril samples formed at 0.1 mg/mL (dotted line), 0.3 mg/mL (solid line), and 0.5 mg/mL (dashed line). Adapted with permission from Yang et al. (2012).
250
Yee-Foong Mok et al.
distributions for fibrils formed at different starting concentrations. These data formed the basis for the development of an isodesmic self-association model coupled to closed-loop formation to describe the formation of apoC-II fibrils (Yang et al., 2012).
3.2 Sedimentation Velocity Procedures Procedures for the analysis of sedimentation data for amyloid fibrils have previously been reviewed (Mok & Howlett, 2006; Pham, Mok, & Howlett, 2009). Briefly, there are a number of considerations. It is important to ensure, when comparing data for amyloid fibrils, that sedimentation experiments are performed under the same conditions of temperature, buffer density, and viscosity. This can be particularly relevant for amyloid fibrils, which often have different structural and solution properties at different temperatures. Experimentally, rotors should be preequilibrated with ultracentrifuge cells assembled ahead of time. The rotor and cells can be subjected to up to 2 h of temperature equilibration in the ultracentrifuge prior to analysis. A recent benchmarking and validation exercise has confirmed the stability of the temperature within the cells when the ultracentrifuge chamber temperature is at 20 °C (Zhao et al., 2014). A further consideration is that large aggregates, such as amyloid fibrils and inclusions, may sediment significantly at 3000 rpm. Thus, minimizing the setup period where the rotor is spinning but data not yet being collected is particularly important. Sedimentation data can be collected using either optical density, interference, or the fluorescence detection system (FDS). For optical density measurements, a wavelength of 280 nm is commonly used although increased sensitivity can be achieved at a wavelength of 230 nm. At this wavelength, advantage is taken of the high absorbance of peptide bonds and high intensity of the light source. Measurements can also be made using the interference optics available on the model XL-I. Interference optics offer the advantage that scans can be collected quickly (<10 s) with high accuracy and sensitivity and are useful for solution conditions where there is interference from strongly absorbing ligands or buffer components. The recent development by AVIV Biomedical of the FDS for the Beckman XL-A analytical ultracentrifuge has greatly increased the versatility and sensitivity of AUC in a broad range of contexts. In particular, the FDS increases the accessible range of analyte concentrations, allowing accurate detection of sedimenting molecules in the picomolar concentration range, and provides specific detection of fluorescent molecules within complex mixtures (Kroe & Laue, 2009). Using this detection system, initiation, laser lock, and setup of the emission gain and experiment settings require the rotor to be spinning
Sedimentation Velocity Analysis of Amyloid Fibrils
251
at 3000 rpm, processes that can take up to 10–15 min to complete. In order to minimize the time of this setup period where fibrils may already start to sediment, fluorescence gains may be set using prior knowledge of the approximate fluorescence yields of samples. Previously saved experimental method conditions may also be used. In addition, firing of the laser requires a centrifuge chamber vacuum below 150 μm. It is thus recommended that the vacuum be allowed to equilibrate to below this value while the rotor is stationary.
4. HETEROGENEOUS SYSTEMS 4.1 Effects of αB-Crystallin on Mature ApoC-II Amyloid Fibrils Continuous sedimentation coefficient (c(s)) distribution analysis can be extended to determine the effects of extrinsic molecules and proteins on the interactions between fibrils. We studied the interaction of the small heat-shock protein, αB-crystallin, with mature apoC-II amyloid fibrils and showed that αB-crystallin binds directly to apoC-II fibrils with an apparent Kd of 5.4 μM (Binger et al., 2013). αB-crystallin exists in solution as a polydisperse and dynamic population of oligomeric complexes, including high-molecular-weight oligomers consisting of more than 50 subunits (Baldwin, Lioe, Robinson, Kay, & Benesch, 2011). Thus, we used analytical ultracentrifugation to investigate the possibility that αB-crystallin oligomers can bind multiple apoC-II fibrils simultaneously thereby mediating their lateral association. Mature apoC-II fibrils formed at an apoC-II concentration of 33 μM were incubated in the presence of increasing concentrations of αBcrystallin up to 15 μM for 72 h and subjected to sedimentation velocity analysis. Sedimentation velocity data (Fig. 4) revealed a significant increase in fibril sedimentation rate in the presence of αB-crystallin at a concentration of 5 μM or greater, indicated by the increased distance between radial scans in these samples. Analysis of the sedimentation velocity data using the c(s) model and conversion of the resulting distributions to weight-average sedimentation coefficients (Sav) confirmed this large increase in sedimentation rate. The calculated weight-average sedimentation coefficient in the presence of 15 μM αB-crystallin was 14-fold that of apoC-II fibrils alone, suggesting that the observed changes in sedimentation behavior were not simply due to binding of αB-crystallin to individual fibrils. Transmission electron microscopy confirmed the presence of large aggregates consisting of many fibrils that were not seen in samples of apoC-II alone, indicating that αB-crystallin induces the lateral association and tangling of apoC-II fibrils.
252
Yee-Foong Mok et al.
Figure 4 Incubation of preformed fibrils with different concentrations of αB-crystallin. Fibrils were incubated with increasing concentrations of αB-crystallin for 72 h and examined by sedimentation velocity and TEM. (A) Radial absorbance scans at 285 nm (circles) of preformed apoC-II fibrils incubated with 0, 0.1, 1, 2, 5, 10, or 15 μM αB-crystallin for 72 h. Fit of the data to the c(s) model is also shown (solid lines). Inset: Sav for each sample. (B) TEM of preformed apoC-II fibrils incubated with 15 μM αBcrystallin for 72 h. Also shown are control samples of apoC-II fibrils incubated for the same period without αB-crystallin (smaller panels). Scale bars correspond to 200 nm. Adapted with permission from Binger et al. (2013).
4.2 The Use of Fluorescence Detection We have recently reviewed the applications of fluorescence detected (FD) AUC to the analysis of amyloid systems, including investigation of amyloidforming proteins labeled with small-molecule fluorophores or green
Sedimentation Velocity Analysis of Amyloid Fibrils
253
Figure 5 Fluorescence-detected sedimentation analysis of apoC-II amyloid fibrils under dilution in the presence and absence of αB-crystallin. (A) Continuous sedimentation coefficient distributions of apoC-II fibrils undiluted or diluted 1:10 and incubated with 0, 2, or 5 μM αB-crystallin for 0 h (solid), 8 h (dotted), 24 h (dashed), or 72 h (dash–dot– dot). Total integrated signal for each distribution was normalized with respect to the corresponding t ¼ 0 sample. Adapted with permission from Binger et al. (2013).
fluorescent protein, and the use of extrinsic fluorophores and fluorescently labeled lipids that bind to amyloid aggregates and intermediates (Mok et al., 2011). We further exploited the ability of FD-AUC to detect the sedimentation of fluorescent components of multicomponent protein mixtures to accurately determine the effect of αB-crystallin on the size distribution of apoC-II fibrils after dilution. Fluorescent apoC-II was prepared by labeling of the protein using Alexa-488 maleimide at an engineered cysteine residue (Ryan, Howlett, & Bailey, 2008), and fibrils were formed from a mixture of 2% Alexa-488-apoC-II and 98% unlabeled, wild-type apoC-II at a total apoC-II concentration of 33 μM. Fluorescent fibrils were then incubated with 0, 2, or 5 μM αB-crystallin and subsequently diluted 1:10 in buffer containing the appropriate concentration of αB-crystallin. After this dilution, the apoC-II concentration was approximately 3.3 μM and, thus, αBcrystallin in the samples comprised a large proportion of the total protein concentration. As high-molecular-weight αB-crystallin oligomers sediment under the centrifugal fields used for analysis of apoC-II fibrils, it was critical to monitor only the fibrillar component of the mixture in order to avoid interference from αB-crystallin and to obtain accurate size distributions and weight-average sedimentation coefficients for the apoC-II fibrils. Analysis of the FD sedimentation velocity data (Fig. 5) showed that dilution of the fibrils in the absence of αB-crystallin resulted in a reduction in the
254
Yee-Foong Mok et al.
weight-average sedimentation coefficient corresponding to fragmentation of fibrils (Binger et al., 2008) and consistent with our model of apoC-II fibril formation including fibril breakage and rejoining (Binger et al., 2008). The presence of αB-crystallin during and after dilution protected fibrils against fragmentation. Interestingly, the sedimentation velocity data also showed a time-dependent reduction in fluorescence signal sedimenting in the presence of αB-crystallin suggesting dissociation of fibrils. Subsequent fibril pelleting assays, performed as described in our earlier review (Mok & Howlett, 2006), confirmed that αB-crystallin mediated the dissociation of fibrils resulting in monomeric apoC-II.
5. SUMMARY The preparative ultracentrifuge provides a useful method for the fractionation of amyloid oligomers, closed loop and linear fibrils, as well as fibrils with different morphologies. Fibril aggregates with sedimentation coefficients up to 10,000 S can be fractionated using this approach. The analytical ultracentrifuge offers much higher precision in the characterization of the size distributions of amyloid fibrils. For fibrils where there is prior knowledge of the relationship between the molecular weight of the fibrils and their diffusion coefficients, c(s) analysis allows the size dependence of fibril diffusion coefficients to be included in the analysis. The recent development of the FDS for the analytical ultracentrifuge has greatly extended the ability to analyze heterogeneous systems and the effects of specific reagents on the sedimentation behavior of amyloid fibrils.
ACKNOWLEDGMENTS We thank Danny Hatters and Angelique Ormsby for permission to include details of the preparation of huntingtin aggregates. This work was supported by the Australian Research Council (projects DP0877800 and DP0984565). M.D.W.G. is the recipient of the C.R. Roper Fellowship and an Australian Research Council Future Fellowship (project number FT140100544).
REFERENCES Antoni, G., Lubberink, M., Estrada, S., Axelsson, J., Carlson, K., Lindsjo, L., et al. (2013). In vivo visualization of amyloid deposits in the heart with 11C-PIB and PET. Journal of Nuclear Medicine: Official Publication, Society of Nuclear Medicine, 54, 213–220. Baldwin, A. J., Lioe, H., Robinson, C. V., Kay, L. E., & Benesch, J. L. (2011). alphaBcrystallin polydispersity is a consequence of unbiased quaternary dynamics. Journal of Molecular Biology, 413, 297–309.
Sedimentation Velocity Analysis of Amyloid Fibrils
255
Binger, K. J., Ecroyd, H., Yang, S., Carver, J. A., Howlett, G. J., & Griffin, M. D. (2013). Avoiding the oligomeric state: AlphaB-crystallin inhibits fragmentation and induces dissociation of apolipoprotein C-II amyloid fibrils. FASEB Journal: Official Publication of the Federation of American Societies for Experimental Biology, 27, 1214–1222. Binger, K. J., Pham, C. L., Wilson, L. M., Bailey, M. F., Lawrence, L. J., Schuck, P., et al. (2008). Apolipoprotein C-II amyloid fibrils assemble via a reversible pathway that includes fibril breaking and rejoining. Journal of Molecular Biology, 376, 1116–1129. Bruggink, K. A., Muller, M., Kuiperij, H. B., & Verbeek, M. M. (2012). Methods for analysis of amyloid-beta aggregates. Journal of Alzheimer’s Disease: JAD, 28, 735–758. Chatani, E., Lee, Y. H., Yagi, H., Yoshimura, Y., Naiki, H., & Goto, Y. (2009). Ultrasonication-dependent production and breakdown lead to minimum-sized amyloid fibrils. Proceedings of the National Academy of Sciences of the United States of America, 106, 11119–11124. Conway, K. A., Lee, S. J., Rochet, J. C., Ding, T. T., Williamson, R. E., & Lansbury, P. T., Jr. (2000). Acceleration of oligomerization, not fibrillization, is a shared property of both alpha-synuclein mutations linked to early-onset Parkinson’s disease: Implications for pathogenesis and therapy. Proceedings of the National Academy of Sciences of the United States of America, 97, 571–576. de Messieres, M., Huang, R. K., He, Y., & Lee, J. C. (2014). Amyloid triangles, squares, and loops of apolipoprotein C-III. Biochemistry, 53, 3261–3263. Demeler, B., & Saber, H. (1998). Determination of molecular parameters by fitting sedimentation data to finite-element solutions of the Lamm equation. Biophysical Journal, 74, 444–454. Fujita, H. (1962). Mathematical theory of sedimentation analysis. New York: Academic Press. Gosal, W. S., Morten, I. J., Hewitt, E. W., Smith, D. A., Thomson, N. H., & Radford, S. E. (2005). Competing pathways determine fibril morphology in the self-assembly of beta2microglobulin into amyloid. Journal of Molecular Biology, 351, 850–864. Griffin, M. D., Mok, M. L., Wilson, L. M., Pham, C. L., Waddington, L. J., Perugini, M. A., et al. (2008). Phospholipid interaction induces molecular-level polymorphism in apolipoprotein C-II amyloid fibrils via alternative assembly pathways. Journal of Molecular Biology, 375, 240–256. Hammarstrom, P., Jiang, X., Deechongkit, S., & Kelly, J. W. (2001). Anion shielding of electrostatic repulsions in transthyretin modulates stability and amyloidosis: Insight into the chaotrope unfolding dichotomy. Biochemistry, 40, 11453–11459. Hatters, D. M., MacPhee, C. E., Lawrence, L. J., Sawyer, W. H., & Howlett, G. J. (2000). Human apolipoprotein C-II forms twisted amyloid ribbons and closed loops. Biochemistry, 39, 8276–8283. Kroe, R. R., & Laue, T. M. (2009). NUTS and BOLTS: applications of fluorescencedetected sedimentation. Analytical Biochemistry, 390, 1–13. Lashuel, H. A., Hartley, D. M., Balakhaneh, D., Aggarwal, A., Teichberg, S., & Callaway, D. J. (2002). New class of inhibitors of amyloid-beta fibril formation. Implications for the mechanism of pathogenesis in Alzheimer’s disease. The Journal of Biological Chemistry, 277, 42881–42890. Lashuel, H. A., Lai, Z., & Kelly, J. W. (1998). Characterization of the transthyretin acid denaturation pathways by analytical ultracentrifugation: Implications for wild-type, V30M, and L55P amyloid fibril formation. Biochemistry, 37, 17851–17864. MacRaild, C. A., Hatters, D. M., Lawrence, L. J., & Howlett, G. J. (2003). Sedimentation velocity analysis of flexible macromolecules: Self-association and tangling of amyloid fibrils. Biophysical Journal, 84, 2562–2569. Mok, Y. F., & Howlett, G. J. (2006). Sedimentation velocity analysis of amyloid oligomers and fibrils. Methods in Enzymology, 413, 199–217.
256
Yee-Foong Mok et al.
Mok, Y. F., Ryan, T. M., Yang, S., Hatters, D. M., Howlett, G. J., & Griffin, M. D. (2011). Sedimentation velocity analysis of amyloid oligomers and fibrils using fluorescence detection. Methods, 54, 67–75. Nagel-Steger, L., Demeler, B., Meyer-Zaika, W., Hochdorffer, K., Schrader, T., & Willbold, D. (2010). Modulation of aggregate size- and shape-distributions of the amyloid-beta peptide by a designed beta-sheet breaker. European Biophysics Journal: EBJ, 39, 415–422. Olshina, M. A., Angley, L. M., Ramdzan, Y. M., Tang, J., Bailey, M. F., Hill, A. F., et al. (2010). Tracking mutant huntingtin aggregation kinetics in cells reveals three major populations that include an invariant oligomer pool. The Journal of Biological Chemistry, 285, 21807–21816. Ormsby, A. R., Ramdzan, Y. M., Mok, Y. F., Jovanoski, K. D., & Hatters, D. M. (2013). A platform to view huntingtin exon 1 aggregation flux in the cell reveals divergent influences from chaperones hsp40 and hsp70. The Journal of Biological Chemistry, 288, 37192–37203. Petkova, A. T., Leapman, R. D., Guo, Z., Yau, W. M., Mattson, M. P., & Tycko, R. (2005). Self-propagating, molecular-level polymorphism in Alzheimer’s beta-amyloid fibrils. Science, 307, 262–265. Pham, C., Mok, Y.-F., & Howlett, G. J. (2009). Sedimentation velocity analysis of amyloid fibrils. Methods in Molecular Biology, 752, 179–196. Ryan, T. M., Howlett, G. J., & Bailey, M. F. (2008). Fluorescence detection of a lipidinduced tetrameric intermediate in amyloid fibril formation by apolipoprotein C-II. The Journal of Biological Chemistry, 283, 35118–35128. Schuck, P. (1998). Sedimentation analysis of noninteracting and self-associating solutes using numerical solutions to the Lamm equation. Biophysical Journal, 75, 1503–1512. Schuck, P. (2000). Size-distribution analysis of macromolecules by sedimentation velocity ultracentrifugation and Lamm equation modeling. Biophysical Journal, 78, 1606–1619. Sipe, J. D., & Cohen, A. S. (2000). Review: History of the amyloid fibril. Journal of Structural Biology, 130, 88–98. Slow, E. J., Graham, R. K., & Hayden, M. R. (2006). To be or not to be toxic: Aggregations in Huntington and Alzheimer disease. Trends in Genetics: TIG, 22, 408–411. Thorn, D. C., Meehan, S., Sunde, M., Rekas, A., Gras, S. L., MacPhee, C. E., et al. (2005). Amyloid fibril formation by bovine milk kappa-casein and its inhibition by the molecular chaperones alphaS- and beta-casein. Biochemistry, 44, 17027–17036. Yang, S., Griffin, M. D., Binger, K. J., Schuck, P., & Howlett, G. J. (2012). An equilibrium model for linear and closed-loop amyloid fibril formation. Journal of Molecular Biology, 421, 364–377. Zerovnik, E. (2002). Amyloid-fibril formation. Proposed mechanisms and relevance to conformational disease. European Journal of Biochemistry/FEBS, 269, 3362–3371. Zhao, H., Balbo, A., Metger, H., Clary, R., Ghirlando, R., & Schuck, P. (2014). Improved measurement of the rotor temperature in analytical ultracentrifugation. Analytical Biochemistry, 451, 69–75.