Biochimica et Biophysica Acta 1844 (2014) 1881–1888
Contents lists available at ScienceDirect
Biochimica et Biophysica Acta journal homepage: www.elsevier.com/locate/bbapap
Elongation of amyloid fibrils through lateral binding of monomers revealed by total internal reflection fluorescence microscopy Hisashi Yagi 1,2, Yuki Abe 1, Naoto Takayanagi, Yuji Goto ⁎ Institute for Protein Research, Osaka University, 3-2 Yamadaoka, Suita, Osaka 565-0871, Japan
a r t i c l e
i n f o
Article history: Received 19 March 2014 Received in revised form 13 June 2014 Accepted 17 June 2014 Available online 12 August 2014 Keywords: Amyloid fibril Islet amyloid polypeptide Surface diffusion Total internal reflection fluorescence microscopy Type II diabetes
a b s t r a c t Amyloid fibrils are fibrillar aggregates of denatured proteins associated with a large number of amyloidoses. The formation of amyloid fibrils has been considered to occur by nucleation and elongation. Real-time imaging of the elongation as well as linear morphology of amyloid fibrils suggests that all elongation events occur at the growing ends of fibrils. On the other hand, we suggested that monomers also bind to the lateral sides of preformed fibrils during the seed-dependent elongation, diffuse to the growing ends, and finally make further conformation changes to the mature amyloid fibrils. To examine lateral binding during the elongation of fibrils, we used islet amyloid polypeptide (IAPP), which has been associated with type II diabetes, and prepared IAPP modified with the fluorescence dye, Alexa532. By monitoring the elongation process with amyloid specific thioflavin T and Alexa532 fluorescence, we obtained overlapping images of the two fluorescence probes, which indicated lateral binding. These results are similar to the surface diffusion-dependent growth of crystals, further supporting the similarities between amyloid fibrillation and the crystallization of substances. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Amyloid fibrils are fibrillar aggregates of denatured proteins with a diameter of approximately 10 nm and length of several μm, and are predominantly composed of cross β-structures in which the direction of polypeptide chains are perpendicular to the fibril axis [1,2]. Amyloid fibrils are associated with a large number of amyloidoses and more than 30 amyloidogenic peptides or proteins have been found including amyloid β peptide (Aβ) associated with Alzheimer's disease, islet amyloid polypeptide (IAPP) with type II diabetes, and β2-microglobulin (β2m) with dialysis-related amyloidosis. On the other hand, peptides or proteins not directly related to diseases have been shown to form amyloid fibrils in vitro under certain conditions [1]. Moreover, various amyloid-like structures formed under physiological conditions, including curli fibrils [3], chorion proteins [4], some peptide hormones [5], and type I antifreeze protein [6], are proposed to be functionally important. Therefore,
Abbreviations: AFM, atomic force microscopy; Aβ, amyloid β; β2m, β2-microglobulin; DMSO, dimethylsulfoxide; FRET, fluorescence resonance energy transfer; HFIP, 1,1,1,3,3,3hexafluoroisopropanol; HPLC, high performance liquid chromatography; IAPP, islet amyloid polypeptide; ThT, thioflavin T; TIRFM, total internal reflection fluorescence microscopy ⁎ Corresponding author. Tel.: +81 6 6879 8614; fax: +81 6 6879 8616. E-mail address: [email protected] (Y. Goto). 1 Both authors contributed equally to this work. 2 Present address: Department of Chemistry and Biotechnology, Graduate School of Engineering and Center for Research on Green Sustainable Chemistry, Tottori University, 4-101 Koyama-cho minami, Tottori, Tottori 680-8552, Japan.
http://dx.doi.org/10.1016/j.bbapap.2014.06.014 1570-9639/© 2014 Elsevier B.V. All rights reserved.
clarifying the mechanism of amyloid fibrillation is of critical importance not only for developing strategies against those diseases, but also for advancing our understanding of proteins. The formation of amyloid fibrils is generally considered to occur by nucleation and elongation [1,7–9]. The nucleation step, in which several monomers associate, is thermodynamically unfavorable. Once the nucleus is formed, elongation proceeds rapidly. Consequently, the kinetics of the spontaneous formation of amyloid fibrils is represented by a sigmoidal curve with a lag phase. The addition of pre-formed fibrils as seeds into the reaction mixtures shortened the lag phase. These phenomena are similar to the crystallization of substances, which indicates that the formation of amyloid fibrils and crystals share common mechanisms even though their morphologies are distinct [7,10]. The linear morphology of amyloid fibrils suggests that all elongation events occur at the growing ends of fibrils, although the distinct roles of the two ends are still unclear [11]. In a template-dependent dock–lock mechanism, the first phase, “dock”, addition of monomers to the growing ends of amyloid template is followed by the second phase, “lock”, stabilization of the deposited peptides [12,13]. Single fibril observations of elongation have also suggested that the growing ends are the only place where monomers or oligomers bind to fibrils [14–16]. On the other hand, the branching of fibrils has been suggested to occur by the formation of a new growing end at the inside of preformed fibrils. Branching is one type of secondary nucleation, in addition to that produced by breaking preformed fibrils [17–20]. On the basis of the kinetic analysis of elongation monitored by tryptophan fluorescence [21] and NMR [22,23], we suggested that monomers also bind to the lateral sides of preformed fibrils. These
1882
H. Yagi et al. / Biochimica et Biophysica Acta 1844 (2014) 1881–1888
molecules that bound to the non-growing ends subsequently diffuse on the lateral surface, reach the growing ends, and finally induce conformation changes to form mature fibrils. Since amyloid formation is the result of the exclusion of peptides or proteins from solvent above their solubility limits [24], such a two-step elongation through lateral binding and diffusion on the surface is likely to occur. “Surface diffusion” of the bound monomers on a terrace (corresponding to the lateral surface) before growth at the step or kink (corresponding to the growing ends) is an important elementary step in the crystallization of substances [25]. Similar non-specific binding and surface diffusion have been proposed as important for the interaction between DNA binding proteins and DNA [26–28]. We used IAPP to examine the possibility of lateral binding during the elongation of fibrils (Fig. 1). IAPP [29], also called amylin [30], is a 37-residue peptide with a molecular weight of 3.9 kDa. The COOHterminal residue is amidated and one intramolecular disulfide bond exists between Cys2 and Cys7. IAPP assumes a dominantly random coil conformation in solution [17]. However, CD and NMR studies have shown that IAPP forms a transient amphipathic helix in the NH2-terminal region upon binding to micelles or membranes [31–33]. This helical intermediate has been proposed to be important for the formation of amyloid fibrils [34]. On the other hand, the COOHterminal region is unstructured. The amyloid fibrils of IAPP were shown to be deposited in the extracellular spaces of the pancreas in patients with type II diabetes [35]. Since amyloid deposits have been observed in approximately 95% of diabetic patients, but are rarely found in non-diabetic individuals, it has been hypothesized that IAPP fibrillation is involved in the pathogenic development of the disease [36,37]. To observe the amyloid fibrils of IAPP, we used total internal reflection fluorescence microscopy (TIRFM). We previously developed a technique for the real time observation of amyloid fibrils at a single fibril level by combining TIRFM and amyloid-specific thioflavin T (ThT) fluorescence [14–16,38]. The direct observation of amyloid fibrils provided a range of important information such as the growth rate, morphology, and elongation direction of amyloid fibrils in real time at the single fibril level. Regarding glucagon, the frequent branching of fibrils during elongation has been observed [20]. If another fluorescence dye could be used in addition to ThT fluorescence, it may be possible to distinguish the molecules in preformed fibrils and intermediate molecules in the process of elongation. Here, we prepared IAPP modified with a fluorescence dye, Alexa532. By monitoring the process of elongation with ThT and Alexa532 fluorescence separately, we succeeded in obtaining overlapping images of the two fluorescence probes. The images obtained indicated the lateral binding of monomers before the formation of mature fibrils, suggesting that surface diffusion common to crystallization is also valid for the formation of amyloid fibrils. 2. Materials and methods 2.1. Materials IAPP peptides were purchased from the Peptide Institute, Inc. (Japan). ThT was obtained from Wako (Japan). Buffer, salts, and solvents were obtained from Nacalai Tesque Co. Ltd. (Japan). Alexa Flour 532 carboxylic acid succinimidyl ester was purchased from Life Technologies (U.S.A.). 2.2. Amyloid fibrils of IAPP Lyophilized synthetic IAPP was dissolved in 10 mM HCl containing 80% (v/v) 1,1,1,3,3,3-hexafluoroisopropanol (HFIP) for complete dissolution. This approach efficiently removed any preformed IAPP aggregates. After lyophilization to remove the HFIP, the peptide was dissolved in 10 mM HCl (pH 2.0) at 100 μM as a stock solution. The
Fig. 1. Primary sequences of human IAPP. IAPP contains disulfide bonds between 2 and 7. All peptides are C-terminally amidated.
formation of fibrils was initiated by diluting the stock solution into the reaction buffer. The final conditions were 25 μM IAPP in 10 mM HCl containing 10% (v/v) HFIP. The peptide solutions were incubated at 25 °C. 2.3. Peptide labeling Succinimidyl esters react with non-protonated aliphatic amine groups. IAPP has two amine groups; the N-terminal α-amino group and ε-amino group of the lysine residue. To specifically label the αamino group, the reaction was carried out at pH 8.0 at which the α-amino group deprotonated while the ε-amino group did not. IAPP was dissolved in 90% (v/v) dimethylsulfoxide (DMSO) to a concentration of 250 μM and was diluted to 100 μM with 20 mM HEPES buffer (pH 8.0) containing 380 μM Alexa Fluor 532 carboxylic acid succinimidyl ester and 0.1 mM EDTA. The solution was incubated for 17 h at room temperature in the dark under gentle stirring. The solution was then lyophilized, dissolved in 80 μl of 80% (v/v) HFIP, and then purified by HPLC. The molecular mass of IAPP-Alexa532 was confirmed by MALDI-TOF-MS (Bruker Daltonics, Germany). 2.4. Reversed phase HPLC The purification of IAPP-Alexa532 was performed on a high performance liquid chromatograph (HPLC) (Gilson Inc., U.S.A.) equipped with COSMOSIL PACKED COLUMN Cholester (4.6 × 250 mm, Nacalai Tesque, Japan). The linear gradient was achieved by a water–acetonitrile system containing 0.05% trifluoroacetic acid with the acetonitrile increment from 20 to 60% for 45 min at 0.5 ml/min. The injection volume of the sample was 10 μl. 2.5. Atomic force microscopy (AFM) Atomic force microscopy (AFM) images were acquired with a Digital Instruments Nanoscope IIIa scanning microscope at room temperature (Bruker AXS, Japan). A 5 μl sample solution was put on freshly cleaved mica. After air drying, measurements were performed in an airtapping mode. The scan rate was 0.5 Hz, and images were obtained in a 5.0 × 5.0 μm area with 512 × 512 points. 2.6. TIRFM The TIRFM system used to observe amyloid fibrils was developed based on an inverted microscope (IX70, Olympus, Japan) as previously described [39]. The ThT molecule was excited at 442 nm by a helium–cadmium (He–Cd) laser (IK5522R-F, Kimmon, Japan). IAPP-Alexa532 was excited by an Argon laser (Spectra-Physics, Japan). The laser power was 20–60 milliwatts (mW) (Argon laser: 20 mW, He–Cd laser: 40–60 mW), and observation period was 0.2–3 s. The fluorescence image was filtered with a bandpass filter (Argon laser: U-MWIGA3, Olympus, Japan, He–Cd laser: D490/30, Omega Optical, U.S.A.) and visualized using a digital steel camera (DP70, Olympus, Japan).
H. Yagi et al. / Biochimica et Biophysica Acta 1844 (2014) 1881–1888
IAPP was added at a final concentration of 10 μM in 10 mM HCl containing 10% HFIP for the real-time observation of IAPP amyloid fibril growth. The ThT solution was added at a final concentration of 5 μM. The sample mixture was placed on a quartz slide, and amyloid fibrils were observed at 25 °C. 3. Results 3.1. Fibrils of IAPP and IAPP-Alexa532 observed by AFM First, we examined the formation of the amyloid fibrils of IAPP in 10 mM HCl at 25 °C in the presence of various concentrations of HFIP monitored by the ThT assay. The optimal HFIP concentration was approximately 10% (v/v), which was consistent with our previous study [23]. On the other hand, the HFIP concentration-dependence of the formation of the fibrils of IAPP-Alexa532 showed a maximum at approximately 30% (v/v), at which intact IAPP forms an α-helical structure. IAPP-Alexa532 was more likely to form amorphous aggregates at 10% HFIP as well as short fibrils of several μm monitored by TIRFM (see Fig. 3B below). These results indicated that IAPP became more hydrophobic with the introduction of Alexa532 and the optimal HFIP concentration to form the fibrils shifted to a higher concentration. We examined the fibrils of IAPP and IAPP-Alexa532 formed under optimal conditions of HFIP using AFM (Fig. 2). AFM images of the two types of fibrils were similar with lengths of several μm and heights of 4–6 nm (Fig. 2A and B). 3.2. Fibrils of IAPP and IAPP-Alexa532 observed by TIRFM We then observed these fibrils using TIRFM (Fig. 3). IAPP fibrils were observed by ThT fluorescence with an excitation at 442 nm and emission at 485 nm (Fig. 3A). The long fibrils that formed on the quartz glass occasionally exhibited branching (Fig. 3A, arrows). On the other hand, the fibrils of IAPP-Alexa532 were observed with an excitation at 514 nm and emission at 554 nm. TIRFM monitored by Alexa532 fluorescence revealed IAPP amyloid fibrils in 10% (v/v) HFIP (Fig. 3B), at which the short fibrils were dominant and amorphous aggregates also existed. The increased hydrophobicity caused by Alexa532 labeling may have led to amorphous aggregates. In contrast, TIRFM images of IAPPAlexa532 fibrils formed in 30% HFIP showed long amyloid fibrils similar to those of non-labeled IAPP (Fig. 3C). As with intact IAPP fibrils, some fibrils of IAPP-Alexa532 exhibited branching at both HFIP concentrations (Fig. 3B and C, arrows).
1883
3.3. Real-time observation of IAPP amyloid fibril growth To obtain further information, we performed real-time observations of fibril growth in the presence of 10% (v/v) HFIP at acidic pH (Fig. 4). Short IAPP fibrils were observed at 4 min and elongated at 20 min. Many fibrils crowded over the plate at 300 min. To compare the fibrils present at 150 min and 270 min, we merged the images of the two time periods colored green and red, respectively (Fig. 5). In this merged image, yellow fibrils were those that existed at 150 min, and red fibrils were newly elongated at 150–270 min. The merged image showed that new IAPP amyloid fibrils elongated from the ends of existing fibrils (Fig. 5-1), which suggested the presence of polarity in the growth direction. Interestingly, some red colors overlapped at the middle parts of preexisting fibrils (Fig. 5-2, and -3), suggesting that IAPP-Alexa532 not only interacted with growing ends, but also interacted with the lateral sides of preformed fibrils. Furthermore, branching of fibrils was detected (Fig. 5-3, and -4). These results also supported the hypothesis that IAPP monomers interact with the lateral sides of amyloid fibrils.
3.4. Observation with two-types of fluorescence dyes We used two types of fluorescence dyes with distinct wavelengths of fluorescence, i.e., ThT and Alexa532, to visualize the interactions between preformed fibrils and monomers (Fig. 6). Because non-labeled and Alexa-532-labeled IAPP formed similar AFM morphologies in the presence of HFIP (Fig. 2), it was difficult to distinguish them by AFM. On the other hand, TIRFM observation may provide images to address the mechanism of fibrillation at the single fibril level. The experiments were performed in 10 mM HCl and 10% (v/v) HFIP because both intact IAPP and Alexa532-labeled IAPP form amyloid fibrils easily. First, we prepared IAPP amyloid fibrils under the same solvent conditions and these fibrils were used as seeds. We then added IAPP-Alexa532 monomers to the IAPP seeds and observed the growth of amyloid fibrils. To precisely analyze the interactions, we changed the colors; green was for the fluorescence of ThT (originally blue) (Fig. 6A), and red was for the fluorescence of Alexa532 (originally orange) (Fig. 6B). We then merged these images (Fig. 6C). IAPPAlexa532 amyloid fibrils were expected to be yellow because green from ThT and red from Alexa532 were mixed. We first confirmed from the images obtained that IAPP-Alexa532 amyloid fibrils elongated from IAPP amyloid fibrils (seeds) (Fig. 6C, 1-3). Branching of IAPP-Alexa532 amyloid fibrils was observed
Fig. 2. Observation of IAPP fibrils by AFM. Twenty-five micromolar IAPP in 10 mM HCl and 10% (v/v) HFIP (A) or 25 μM IAPP-Alexa532 in 10 mM HCl and 30% (v/v) HFIP (B) were incubated for 1 day to make fibrils. Scale bars represent 1 μm.
1884
H. Yagi et al. / Biochimica et Biophysica Acta 1844 (2014) 1881–1888
Fig. 3. Amyloid fibrils of IAPP and IAPP-Alexa532 observed by TIRFM. A, fibrils of IAPP prepared by incubating for 1 day in 10 mM HCl and 10% (v/v) HFIP. B, and C, fibrils of IAPP-Alexa532 prepared by incubating for 1 day in 10 mM HCl and 10% (B) or 30% (C) (v/v) HFIP. The arrows show the branching of amyloid fibrils (see expanded images). Scale bars represent 10 μm.
(Fig. 6C-1). We also found that IAPP-Alexa532 amyloid fibrils elongated from both ends of the IAPP amyloid fibrils (Fig. 6C-2). Most importantly, the sides of preformed fibrils were mostly covered with Alexa532 fluorescence, exhibited by a yellow color. Without lateral binding, preformed fibrils should remain green in the merged images. We also show additional examples of overlapped images of Alexa532 and ThT, showing that a large number of preformed fibrils interacted with Alexa532-labeled IAPP (Fig. 6D). Therefore, these results provide strong evidence that IAPP-Alexa532 monomers interacted with the lateral sides of preformed fibrils as well as the elongating sites of amyloid fibrils. We expected IAPP-Alexa532 amyloid fibrils to be yellow, however, some fibrils remained red, which suggests that fluorescence resonance energy transfer (FRET) occurred between ThT and Alexa532. Consequently, ThT fluorescence became weak in the presence of IAPPAlexa532 monomers or fibrils, therefore, some fibrils exhibited a red color in merged images. Because FRET only occurs when the distance between the donor and acceptor molecules is very close, these results
are consistent with the hypothesis that IAPP fibrils and IAPP-Alexa532 monomers interacted with each other at the lateral sides of fibrils. 4. Discussion Surface diffusion at an interface between a solution and a protein crystal was previously shown to be an important elementary process of the crystal growth of substances [25]. Protein molecules form crystals in a supersaturated solution by nucleation followed by growth [40,41]. At above critical concentrations of solubility and in the presence of seed crystals, solute molecules started crystallization by adsorbing to the wide “terrace” of seed crystals (Fig. 7A-1). The bound molecules diffused randomly at the terrace, reaching the “step” and then “kink”, the growing sites of single crystals (Fig. 7A-2). The concentration of bound molecules was significantly higher than the bulk concentration, however, the diffusion constant decreased because of the strong interaction with the crystal surface. The gradual immobilization of bound molecules proceeded at the steps and kinks of crystals with further stabilizing
H. Yagi et al. / Biochimica et Biophysica Acta 1844 (2014) 1881–1888
1885
Fig. 4. Real-time observation of the elongation of IAPP amyloid fibrils. IAPP and ThT concentrations were 10 μM and 5 μM, respectively. Conditions were 10 mM HCl in the presence of 10% (v/v) HFIP at 25 °C. Scale bars represent 10 μm.
interactions, finally leading to the growth of crystals. This mechanism with surface diffusion has been linked to the overall rapid rate of crystallization in the presence of seed crystals in which surface diffusion is a critical step ensuring rapidity.
Considering the similarities between crystallization and the formation of amyloid fibrils [7], it is likely that a similar surface diffusion play important roles in the elongation process of amyloid fibrils (Fig. 7B-1). However, linear morphologies were often rigid and straight,
Fig. 5. Merged images of elongating fibrils at 150 min (green) and 270 min (red). Red amyloid fibrils are elongating fibrils. (1–4) are close-up views of the active sites of elongation. (1–2) shows elongation at both ends of the amyloid fibrils. (3–4) shows the branching of amyloid fibrils. The scale bar represents 10 μm.
1886
H. Yagi et al. / Biochimica et Biophysica Acta 1844 (2014) 1881–1888
Fig. 6. Two-color fluorescence observation of amyloid fibril growth. IAPP-Alexa532 monomers were added to IAPP amyloid fibrils prepared in 10 mM HCl and 10% (v/v) HFIP at 25 °C. A–C, fluorescence images detected by ThT (A), those by Alexa532 (B), and merged images (C). C1–3, IAPP-Alexa532 fibrils (red) elongated from IAPP fibrils (yellow). C1 and 3 show the branching of amyloid fibrils. C2 shows the elongation of IAPP-Alexa532 fibrils (red) from both ends of IAPP fibrils (yellow). (D) Additional examples of overlapped images of Alexa532 and ThT.
which suggested only the importance of the ends of fibrils for elongation (Fig. 7B-2). Real time imaging of fibril elongation also did not indicate the importance of regions other than the growing ends of fibrils. However, detailed analysis of the kinetics of fibril elongation indicated several pieces of evidence inconsistent with the simple growing mechanism, taking only the active ends of the fibrils into account. First, branching of fibrils from the trunk has been observed in several cases [17–20]. Regarding the seed-dependent growth of glucagon fibrils, branching from the preformed fibrils occurred frequently to form the brush-like morphology of fibrils [20]. Kinetic analysis with β2m amyloid fibrils at pH 2.5 monitored by tryptophan fluorescence indicated that the stoichiometry of “seed fibrils”:“bound monomers” was approximately 10:1 [21]. Considering the minimal size of seed fibrils made of several hundreds of monomers, this stoichiometry is too low to make binding sites at the ends of linear fibrils. The low stoichiometry can instead be easily explained by the lateral binding of monomers at the sides of fibrils. The lateral binding of β2m was further supported by an analysis of the intermediate of fibril elongation monitored by NMR [22,23]. We performed various NMR experiments to characterize the interaction of seed fibrils and acid-denatured monomeric β2m. An analysis of
transverse relaxation rates revealed that acid-denatured β2m underwent a structural exchange with an extensively unfolded form. The results suggested that the partially structured acid-denatured β2m was “activated” to become extensively unfolded, in which state the hydrophobic residues were exposed and associated with the seed. The low stoichiometry of seeds and bound monomers again supported the lateral binding model. We proposed a general model of amyloid elongation in comparison with the crystal growth of substances. Amyloid formation could only start under supersaturation above the solubility limit of amyloidogenic peptides. However, because of this supersaturation, it only started after a long lag time. Upon the addition of seed fibrils, fibril formation began immediately taking advantage of template-assisted reactions. As is the case with crystallization, the absorption of monomers occurred at the fibril surfaces. The major driving force of adsorption in amyloid fibrillation is hydrophobic interactions. However, these interactions were not strong enough to fix the bound molecules, therefore, they diffused on the fibril/water interface along with the dissociation/binding processes (Fig. 7B-1). When diffusing molecules reached the growing ends, tighter binding, coupled with conformational changes, occurred to finish one round of fibril growth (Fig. 7B-2). The model shown in
H. Yagi et al. / Biochimica et Biophysica Acta 1844 (2014) 1881–1888
1887
Fig. 7. Comparison of the surface diffusion-dependent formation of crystals and amyloid fibrils. A, schematic model of crystal growth. (1) Monomers interact with the terrace of crystals and undergo surface diffusion as well as dissociation/association. (2) They then bind to the step or kink to form crystals. (3) Alternatively, monomers associate at the terrace to form the island, a new nucleus for crystallization. B, schematic model of amyloid fibril growth. (1) IAPP monomers interact with the lateral sides of fibrils. Monomers repeat binding/dissociation and surface diffusion. (2) They then move to the growing ends to elongate fibrils. (3) Alternatively, monomers form the nucleus at the lateral side, and the branching of fibrils occurs.
Fig. 7B satisfactorily explains the mechanism of the branching of fibrils [17–20]. In the case of crystallization, several bound molecules on the same terrace diffused, associated with each other, and nucleated, leading to “island” growth (Fig. 7A-3). A similar association of monomers via surface diffusion may make a secondary nucleus, making branching possible (Fig. 7B-3). In amyloid fibrillation, the strong precipitation conditions were shown to lead to the formation of amorphous aggregates without specific interactions [24]. Amorphous aggregates correspond to glasses in crystallography. The formation of amorphous aggregates can also be explained on the basis of a surface diffusion model. Upon binding of monomers to seed fibrils, very strong interactions prevented the subsequent surface diffusion of bound monomers. Because binding continues to occur, this leads to the accumulation of trapped monomers and, consequently, the formation of amorphous aggregates. The same is true even in the absence of seed fibrils. Thus, lateral binding and surface diffusion are steps required for the formation of ordered amyloid fibrils, as is the case for the formation of single crystals. Finally, the mechanism of fibril elongation including the surface diffusion of laterally bound intermediates resembles the search for some DNA-binding proteins. Intramolecular as well as intermolecular translocation is considered to contribute significantly to the speed of a search for DNA-bound proteins [26–28]. Taken together, a sequential mechanism with weak binding, surface diffusion, and final tight docking may be a common mechanism maximizing the kinetic efficiency achieving highly organized interactions. By distinguishing the seed fibrils and newly formed fibrils of IAPP due to the introduction of two types of fluorescence probes, we succeeded in showing the lateral binding of monomers to preformed fibrils and the branching of fibrils from preformed fibrils. This type of lateral binding is well known to be an important elementary step for the crystallization of substances, i.e. surface diffusion. Surface diffusion determines the growth rate of crystals and their morphologies. Considering the similarities in the formation of amyloid fibrils and crystallization, it is clear that surface diffusion plays a role in the formation of amyloid fibrils, although the morphologies of fibrils
suggest the importance of the growing ends of fibrils. The inability of surface diffusion under very strong protein precipitation conditions leads to the formation of glass-like amorphous aggregates. Taken together, we proposed a new model of amyloid fibril growth in which lateral binding plays an important role. Further studies taking into account the roles of surface diffusion will advance our understanding of amyloid fibrils. Acknowledgements This work was supported by the Ministry of Education, Science, Sports, and Culture, Grant for Scientific Research, 22810014 (2010–2011), 24370067 (2012–2015), and 24592618 (2012–2015), Japan. We acknowledge financial support from Takeda Science Foundation.
References [1] F. Chiti, C.M. Dobson, Protein misfolding, functional amyloid, and human disease, Annu. Rev. Biochem. 75 (2006) 333–366. [2] J.D. Sipe, M.D. Benson, J.N. Buxbaum, S. Ikeda, G. Merlini, M.J. Saraiva, P. Westermark, Amyloid fibril protein nomenclature: 2012 recommendations from the Nomenclature Committee of the International Society of Amyloidosis, Amyloid 19 (2012) 167–170. [3] M.M. Barnhart, M.R. Chapman, Curli biogenesis and function, Annu. Rev. Microbiol. 60 (2006) 131–147. [4] V.A. Iconomidou, G. Vriend, S.J. Hamodrakas, Amyloids protect the silkmoth oocyte and embryo, FEBS Lett. 479 (2000) 141–145. [5] S.K. Maji, M.H. Perrin, M.R. Sawaya, S. Jessberger, K. Vadodaria, R.A. Rissman, P.S. Singru, K.P. Nilsson, R. Simon, D. Schubert, D. Eisenberg, J. Rivier, P. Sawchenko, W. Vale, R. Riek, Functional amyloids as natural storage of peptide hormones in pituitary secretory granules, Science 325 (2009) 328–332. [6] S.P. Graether, C.M. Slupsky, B.D. Sykes, Freezing of a fish antifreeze protein results in amyloid fibril formation, Biophys. J. 84 (2003) 552–557. [7] J.T. Jarrett, P.T. Lansbury Jr., Seeding “one-dimensional crystallization” of amyloid: a pathogenic mechanism in Alzheimer's disease and scrapie? Cell 73 (1993) 1055–1058. [8] H. Naiki, S. Hashimoto, H. Suzuki, K. Kimura, K. Nakakuki, F. Gejyo, Establishment of a kinetic model of dialysis-related amyloid fibril extension in vitro, Amyloid 4 (1997) 223–232.
1888
H. Yagi et al. / Biochimica et Biophysica Acta 1844 (2014) 1881–1888
[9] R. Wetzel, Kinetics and thermodynamics of amyloid fibril assembly, Acc. Chem. Res. 39 (2006) 671–679. [10] P.T. Lansbury Jr., B. Caughey, The chemistry of scrapie infection: implications of the ‘ice 9’ metaphor, Chem. Biol. 2 (1995) 1–5. [11] A.H. DePace, J.S. Weissman, Origins and kinetic consequences of diversity in Sup35 yeast prion fibers, Nat. Struct. Biol. 9 (2002) 389–396. [12] W.P. Esler, E.R. Stimson, J.M. Jennings, H.V. Vinters, J.R. Ghilardi, J.P. Lee, P.W. Mantyh, J.E. Maggio, Alzheimer's disease amyloid propagation by a templatedependent dock–lock mechanism, Biochemistry 39 (2000) 6288–6295. [13] P.H. Nguyen, M.S. Li, G. Stock, J.E. Straub, D. Thirumalai, Monomer adds to preformed structured oligomers of Aβ-peptides by a two-stage dock–lock mechanism, Proc. Natl. Acad. Sci. U. S. A. 104 (2007) 111–116. [14] T. Ban, M. Hoshino, S. Takahashi, D. Hamada, K. Hasegawa, H. Naiki, Y. Goto, Direct observation of Aβ amyloid fibril growth and inhibition, J. Mol. Biol. 344 (2004) 757–767. [15] T. Ban, K. Yamaguchi, Y. Goto, Direct observation of amyloid fibril growth, propagation, and adaptation, Acc. Chem. Res. 39 (2006) 663–670. [16] H. Yagi, T. Ban, K. Morigaki, H. Naiki, Y. Goto, Visualization and classification of amyloid β supramolecular assemblies, Biochemistry 46 (2007) 15009–15017. [17] S.B. Padrick, A.D. Miranker, Islet amyloid polypeptide: identification of long-range contacts and local order on the fibrillogenesis pathway, J. Mol. Biol. 308 (2001) 783–794. [18] F. Librizzi, C. Rischel, The kinetic behavior of insulin fibrillation is determined by heterogeneous nucleation pathways, Protein Sci. 14 (2005) 3129–3134. [19] M. Manno, E.F. Craparo, A. Podesta, D. Bulone, R. Carrotta, V. Martorana, G. Tiana, P.L. San Biagio, Kinetics of different processes in human insulin amyloid formation, J. Mol. Biol. 366 (2007) 258–274. [20] C.B. Andersen, H. Yagi, M. Manno, V. Martorana, T. Ban, G. Christiansen, D.E. Otzen, Y. Goto, C. Rischel, Branching in amyloid fibril growth, Biophys. J. 96 (2009) 1529–1536. [21] E. Chatani, R. Ohnishi, T. Konuma, K. Sakurai, H. Naiki, Y. Goto, Pre-steady-state kinetic analysis of the elongation of amyloid fibrils of β(2)-microglobulin with tryptophan mutagenesis, J. Mol. Biol. 400 (2010) 1057–1066. [22] T. Konuma, E. Chatani, M. Yagi, K. Sakurai, T. Ikegami, H. Naiki, Y. Goto, Kinetic intermediates of β(2)-microglobulin fibril elongation probed by pulse-labeling H/D exchange combined with NMR analysis, J. Mol. Biol. 405 (2011) 851–862. [23] K. Yanagi, K. Sakurai, Y. Yoshimura, T. Konuma, Y.H. Lee, K. Sugase, T. Ikegami, H. Naiki, Y. Goto, The monomer–seed interaction mechanism in the formation of the β2-microglobulin amyloid fibril clarified by solution NMR techniques, J. Mol. Biol. 422 (2012) 390–402. [24] Y. Yoshimura, Y. Lin, H. Yagi, Y.H. Lee, H. Kitayama, K. Sakurai, M. So, H. Ogi, H. Naiki, Y. Goto, Distinguishing crystal-like amyloid fibrils and glass-like amorphous aggregates from their kinetics of formation, Proc. Natl. Acad. Sci. U. S. A. 109 (2012) 14446–14451. [25] G. Sazaki, M. Okada, T. Matsui, T. Watanabe, H. Higuchi, K. Tsukamoto, K. Nakajima, Single-molecule visualization of diffusion at the solution — crystal interface, Cryst. Growth Des. 8 (2008) 2024–2031.
[26] J. Iwahara, G.M. Clore, Detecting transient intermediates in macromolecular binding by paramagnetic NMR, Nature 440 (2006) 1227–1230. [27] M. Doucleff, G.M. Clore, Global jumping and domain-specific intersegment transfer between DNA cognate sites of the multidomain transcription factor Oct-1, Proc. Natl. Acad. Sci. U. S. A. 105 (2008) 13871–13876. [28] P. Lian, L. Angela Liu, Y. Shi, Y. Bu, D. Wei, Tethered-hopping model for protein–DNA binding and unbinding based on Sox2–Oct1–Hoxb1 ternary complex simulations, Biophys. J. 98 (2010) 1285–1293. [29] P. Westermark, C. Wernstedt, E. Wilander, D.W. Hayden, T.D. O'Brien, K.H. Johnson, Amyloid fibrils in human insulinoma and islets of Langerhans of the diabetic cat are derived from a neuropeptide-like protein also present in normal islet cells, Proc. Natl. Acad. Sci. U. S. A. 84 (1987) 3881–3885. [30] G.J. Cooper, A.C. Willis, A. Clark, R.C. Turner, R.B. Sim, K.B. Reid, Purification and characterization of a peptide from amyloid-rich pancreases of type 2 diabetic patients, Proc. Natl. Acad. Sci. U. S. A. 84 (1987) 8628–8632. [31] J.D. Knight, J.A. Hebda, A.D. Miranker, Conserved and cooperative assembly of membrane-bound alpha-helical states of islet amyloid polypeptide, Biochemistry 45 (2006) 9496–9508. [32] S.M. Patil, S. Xu, S.R. Sheftic, A.T. Alexandrescu, Dynamic α-helix structure of micelle-bound human amylin, J. Biol. Chem. 284 (2009) 11982–11991. [33] A. Abedini, D.P. Raleigh, A critical assessment of the role of helical intermediates in amyloid formation by natively unfolded proteins and polypeptides, Protein Eng. Des. Sel. 22 (2009) 453–459. [34] R.P. Nanga, J.R. Brender, S. Vivekanandan, N. Popovych, A. Ramamoorthy, NMR structure in a membrane environment reveals putative amyloidogenic regions of the SEVI precursor peptide PAP(248–286), J. Am. Chem. Soc. 131 (2009) 17972–17979. [35] R.L. Hull, G.T. Westermark, P. Westermark, S.E. Kahn, Islet amyloid: a critical entity in the pathogenesis of type 2 diabetes, J. Clin. Endocrinol. Metab. 89 (2004) 3629–3643. [36] M.F. Engel, L. Khemtemourian, C.C. Kleijer, H.J. Meeldijk, J. Jacobs, A.J. Verkleij, B. de Kruijff, J.A. Killian, J.W. Hoppener, Membrane damage by human islet amyloid polypeptide through fibril growth at the membrane, Proc. Natl. Acad. Sci. U. S. A. 105 (2008) 6033–6038. [37] A. Quist, I. Doudevski, H. Lin, R. Azimova, D. Ng, B. Frangione, B. Kagan, J. Ghiso, R. Lal, Amyloid ion channels: a common structural link for protein-misfolding disease, Proc. Natl. Acad. Sci. U. S. A. 102 (2005) 10427–10432. [38] K. Yanagi, M. Ashizaki, H. Yagi, K. Sakurai, Y.H. Lee, Y. Goto, Hexafluoroisopropanol induces amyloid fibrils of islet amyloid polypeptide by enhancing both hydrophobic and electrostatic interactions, J. Biol. Chem. 286 (2011) 23959–23966. [39] T. Ban, D. Hamada, K. Hasegawa, H. Naiki, Y. Goto, Direct observation of amyloid fibril growth monitored by thioflavin T fluorescence, J. Biol. Chem. 278 (2003) 16462–16465. [40] Z. Kam, H.B. Shore, G. Feher, On the crystallization of proteins, J. Mol. Biol. 123 (1978) 539–555. [41] S.D. Durbin, G. Feher, Protein crystallization, Annu. Rev. Phys. Chem. 47 (1996) 171–204.
Contents lists available at ScienceDirect
Biochimica et Biophysica Acta journal homepage: www.elsevier.com/locate/bbapap
Elongation of amyloid fibrils through lateral binding of monomers revealed by total internal reflection fluorescence microscopy Hisashi Yagi 1,2, Yuki Abe 1, Naoto Takayanagi, Yuji Goto ⁎ Institute for Protein Research, Osaka University, 3-2 Yamadaoka, Suita, Osaka 565-0871, Japan
a r t i c l e
i n f o
Article history: Received 19 March 2014 Received in revised form 13 June 2014 Accepted 17 June 2014 Available online 12 August 2014 Keywords: Amyloid fibril Islet amyloid polypeptide Surface diffusion Total internal reflection fluorescence microscopy Type II diabetes
a b s t r a c t Amyloid fibrils are fibrillar aggregates of denatured proteins associated with a large number of amyloidoses. The formation of amyloid fibrils has been considered to occur by nucleation and elongation. Real-time imaging of the elongation as well as linear morphology of amyloid fibrils suggests that all elongation events occur at the growing ends of fibrils. On the other hand, we suggested that monomers also bind to the lateral sides of preformed fibrils during the seed-dependent elongation, diffuse to the growing ends, and finally make further conformation changes to the mature amyloid fibrils. To examine lateral binding during the elongation of fibrils, we used islet amyloid polypeptide (IAPP), which has been associated with type II diabetes, and prepared IAPP modified with the fluorescence dye, Alexa532. By monitoring the elongation process with amyloid specific thioflavin T and Alexa532 fluorescence, we obtained overlapping images of the two fluorescence probes, which indicated lateral binding. These results are similar to the surface diffusion-dependent growth of crystals, further supporting the similarities between amyloid fibrillation and the crystallization of substances. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Amyloid fibrils are fibrillar aggregates of denatured proteins with a diameter of approximately 10 nm and length of several μm, and are predominantly composed of cross β-structures in which the direction of polypeptide chains are perpendicular to the fibril axis [1,2]. Amyloid fibrils are associated with a large number of amyloidoses and more than 30 amyloidogenic peptides or proteins have been found including amyloid β peptide (Aβ) associated with Alzheimer's disease, islet amyloid polypeptide (IAPP) with type II diabetes, and β2-microglobulin (β2m) with dialysis-related amyloidosis. On the other hand, peptides or proteins not directly related to diseases have been shown to form amyloid fibrils in vitro under certain conditions [1]. Moreover, various amyloid-like structures formed under physiological conditions, including curli fibrils [3], chorion proteins [4], some peptide hormones [5], and type I antifreeze protein [6], are proposed to be functionally important. Therefore,
Abbreviations: AFM, atomic force microscopy; Aβ, amyloid β; β2m, β2-microglobulin; DMSO, dimethylsulfoxide; FRET, fluorescence resonance energy transfer; HFIP, 1,1,1,3,3,3hexafluoroisopropanol; HPLC, high performance liquid chromatography; IAPP, islet amyloid polypeptide; ThT, thioflavin T; TIRFM, total internal reflection fluorescence microscopy ⁎ Corresponding author. Tel.: +81 6 6879 8614; fax: +81 6 6879 8616. E-mail address: [email protected] (Y. Goto). 1 Both authors contributed equally to this work. 2 Present address: Department of Chemistry and Biotechnology, Graduate School of Engineering and Center for Research on Green Sustainable Chemistry, Tottori University, 4-101 Koyama-cho minami, Tottori, Tottori 680-8552, Japan.
http://dx.doi.org/10.1016/j.bbapap.2014.06.014 1570-9639/© 2014 Elsevier B.V. All rights reserved.
clarifying the mechanism of amyloid fibrillation is of critical importance not only for developing strategies against those diseases, but also for advancing our understanding of proteins. The formation of amyloid fibrils is generally considered to occur by nucleation and elongation [1,7–9]. The nucleation step, in which several monomers associate, is thermodynamically unfavorable. Once the nucleus is formed, elongation proceeds rapidly. Consequently, the kinetics of the spontaneous formation of amyloid fibrils is represented by a sigmoidal curve with a lag phase. The addition of pre-formed fibrils as seeds into the reaction mixtures shortened the lag phase. These phenomena are similar to the crystallization of substances, which indicates that the formation of amyloid fibrils and crystals share common mechanisms even though their morphologies are distinct [7,10]. The linear morphology of amyloid fibrils suggests that all elongation events occur at the growing ends of fibrils, although the distinct roles of the two ends are still unclear [11]. In a template-dependent dock–lock mechanism, the first phase, “dock”, addition of monomers to the growing ends of amyloid template is followed by the second phase, “lock”, stabilization of the deposited peptides [12,13]. Single fibril observations of elongation have also suggested that the growing ends are the only place where monomers or oligomers bind to fibrils [14–16]. On the other hand, the branching of fibrils has been suggested to occur by the formation of a new growing end at the inside of preformed fibrils. Branching is one type of secondary nucleation, in addition to that produced by breaking preformed fibrils [17–20]. On the basis of the kinetic analysis of elongation monitored by tryptophan fluorescence [21] and NMR [22,23], we suggested that monomers also bind to the lateral sides of preformed fibrils. These
1882
H. Yagi et al. / Biochimica et Biophysica Acta 1844 (2014) 1881–1888
molecules that bound to the non-growing ends subsequently diffuse on the lateral surface, reach the growing ends, and finally induce conformation changes to form mature fibrils. Since amyloid formation is the result of the exclusion of peptides or proteins from solvent above their solubility limits [24], such a two-step elongation through lateral binding and diffusion on the surface is likely to occur. “Surface diffusion” of the bound monomers on a terrace (corresponding to the lateral surface) before growth at the step or kink (corresponding to the growing ends) is an important elementary step in the crystallization of substances [25]. Similar non-specific binding and surface diffusion have been proposed as important for the interaction between DNA binding proteins and DNA [26–28]. We used IAPP to examine the possibility of lateral binding during the elongation of fibrils (Fig. 1). IAPP [29], also called amylin [30], is a 37-residue peptide with a molecular weight of 3.9 kDa. The COOHterminal residue is amidated and one intramolecular disulfide bond exists between Cys2 and Cys7. IAPP assumes a dominantly random coil conformation in solution [17]. However, CD and NMR studies have shown that IAPP forms a transient amphipathic helix in the NH2-terminal region upon binding to micelles or membranes [31–33]. This helical intermediate has been proposed to be important for the formation of amyloid fibrils [34]. On the other hand, the COOHterminal region is unstructured. The amyloid fibrils of IAPP were shown to be deposited in the extracellular spaces of the pancreas in patients with type II diabetes [35]. Since amyloid deposits have been observed in approximately 95% of diabetic patients, but are rarely found in non-diabetic individuals, it has been hypothesized that IAPP fibrillation is involved in the pathogenic development of the disease [36,37]. To observe the amyloid fibrils of IAPP, we used total internal reflection fluorescence microscopy (TIRFM). We previously developed a technique for the real time observation of amyloid fibrils at a single fibril level by combining TIRFM and amyloid-specific thioflavin T (ThT) fluorescence [14–16,38]. The direct observation of amyloid fibrils provided a range of important information such as the growth rate, morphology, and elongation direction of amyloid fibrils in real time at the single fibril level. Regarding glucagon, the frequent branching of fibrils during elongation has been observed [20]. If another fluorescence dye could be used in addition to ThT fluorescence, it may be possible to distinguish the molecules in preformed fibrils and intermediate molecules in the process of elongation. Here, we prepared IAPP modified with a fluorescence dye, Alexa532. By monitoring the process of elongation with ThT and Alexa532 fluorescence separately, we succeeded in obtaining overlapping images of the two fluorescence probes. The images obtained indicated the lateral binding of monomers before the formation of mature fibrils, suggesting that surface diffusion common to crystallization is also valid for the formation of amyloid fibrils. 2. Materials and methods 2.1. Materials IAPP peptides were purchased from the Peptide Institute, Inc. (Japan). ThT was obtained from Wako (Japan). Buffer, salts, and solvents were obtained from Nacalai Tesque Co. Ltd. (Japan). Alexa Flour 532 carboxylic acid succinimidyl ester was purchased from Life Technologies (U.S.A.). 2.2. Amyloid fibrils of IAPP Lyophilized synthetic IAPP was dissolved in 10 mM HCl containing 80% (v/v) 1,1,1,3,3,3-hexafluoroisopropanol (HFIP) for complete dissolution. This approach efficiently removed any preformed IAPP aggregates. After lyophilization to remove the HFIP, the peptide was dissolved in 10 mM HCl (pH 2.0) at 100 μM as a stock solution. The
Fig. 1. Primary sequences of human IAPP. IAPP contains disulfide bonds between 2 and 7. All peptides are C-terminally amidated.
formation of fibrils was initiated by diluting the stock solution into the reaction buffer. The final conditions were 25 μM IAPP in 10 mM HCl containing 10% (v/v) HFIP. The peptide solutions were incubated at 25 °C. 2.3. Peptide labeling Succinimidyl esters react with non-protonated aliphatic amine groups. IAPP has two amine groups; the N-terminal α-amino group and ε-amino group of the lysine residue. To specifically label the αamino group, the reaction was carried out at pH 8.0 at which the α-amino group deprotonated while the ε-amino group did not. IAPP was dissolved in 90% (v/v) dimethylsulfoxide (DMSO) to a concentration of 250 μM and was diluted to 100 μM with 20 mM HEPES buffer (pH 8.0) containing 380 μM Alexa Fluor 532 carboxylic acid succinimidyl ester and 0.1 mM EDTA. The solution was incubated for 17 h at room temperature in the dark under gentle stirring. The solution was then lyophilized, dissolved in 80 μl of 80% (v/v) HFIP, and then purified by HPLC. The molecular mass of IAPP-Alexa532 was confirmed by MALDI-TOF-MS (Bruker Daltonics, Germany). 2.4. Reversed phase HPLC The purification of IAPP-Alexa532 was performed on a high performance liquid chromatograph (HPLC) (Gilson Inc., U.S.A.) equipped with COSMOSIL PACKED COLUMN Cholester (4.6 × 250 mm, Nacalai Tesque, Japan). The linear gradient was achieved by a water–acetonitrile system containing 0.05% trifluoroacetic acid with the acetonitrile increment from 20 to 60% for 45 min at 0.5 ml/min. The injection volume of the sample was 10 μl. 2.5. Atomic force microscopy (AFM) Atomic force microscopy (AFM) images were acquired with a Digital Instruments Nanoscope IIIa scanning microscope at room temperature (Bruker AXS, Japan). A 5 μl sample solution was put on freshly cleaved mica. After air drying, measurements were performed in an airtapping mode. The scan rate was 0.5 Hz, and images were obtained in a 5.0 × 5.0 μm area with 512 × 512 points. 2.6. TIRFM The TIRFM system used to observe amyloid fibrils was developed based on an inverted microscope (IX70, Olympus, Japan) as previously described [39]. The ThT molecule was excited at 442 nm by a helium–cadmium (He–Cd) laser (IK5522R-F, Kimmon, Japan). IAPP-Alexa532 was excited by an Argon laser (Spectra-Physics, Japan). The laser power was 20–60 milliwatts (mW) (Argon laser: 20 mW, He–Cd laser: 40–60 mW), and observation period was 0.2–3 s. The fluorescence image was filtered with a bandpass filter (Argon laser: U-MWIGA3, Olympus, Japan, He–Cd laser: D490/30, Omega Optical, U.S.A.) and visualized using a digital steel camera (DP70, Olympus, Japan).
H. Yagi et al. / Biochimica et Biophysica Acta 1844 (2014) 1881–1888
IAPP was added at a final concentration of 10 μM in 10 mM HCl containing 10% HFIP for the real-time observation of IAPP amyloid fibril growth. The ThT solution was added at a final concentration of 5 μM. The sample mixture was placed on a quartz slide, and amyloid fibrils were observed at 25 °C. 3. Results 3.1. Fibrils of IAPP and IAPP-Alexa532 observed by AFM First, we examined the formation of the amyloid fibrils of IAPP in 10 mM HCl at 25 °C in the presence of various concentrations of HFIP monitored by the ThT assay. The optimal HFIP concentration was approximately 10% (v/v), which was consistent with our previous study [23]. On the other hand, the HFIP concentration-dependence of the formation of the fibrils of IAPP-Alexa532 showed a maximum at approximately 30% (v/v), at which intact IAPP forms an α-helical structure. IAPP-Alexa532 was more likely to form amorphous aggregates at 10% HFIP as well as short fibrils of several μm monitored by TIRFM (see Fig. 3B below). These results indicated that IAPP became more hydrophobic with the introduction of Alexa532 and the optimal HFIP concentration to form the fibrils shifted to a higher concentration. We examined the fibrils of IAPP and IAPP-Alexa532 formed under optimal conditions of HFIP using AFM (Fig. 2). AFM images of the two types of fibrils were similar with lengths of several μm and heights of 4–6 nm (Fig. 2A and B). 3.2. Fibrils of IAPP and IAPP-Alexa532 observed by TIRFM We then observed these fibrils using TIRFM (Fig. 3). IAPP fibrils were observed by ThT fluorescence with an excitation at 442 nm and emission at 485 nm (Fig. 3A). The long fibrils that formed on the quartz glass occasionally exhibited branching (Fig. 3A, arrows). On the other hand, the fibrils of IAPP-Alexa532 were observed with an excitation at 514 nm and emission at 554 nm. TIRFM monitored by Alexa532 fluorescence revealed IAPP amyloid fibrils in 10% (v/v) HFIP (Fig. 3B), at which the short fibrils were dominant and amorphous aggregates also existed. The increased hydrophobicity caused by Alexa532 labeling may have led to amorphous aggregates. In contrast, TIRFM images of IAPPAlexa532 fibrils formed in 30% HFIP showed long amyloid fibrils similar to those of non-labeled IAPP (Fig. 3C). As with intact IAPP fibrils, some fibrils of IAPP-Alexa532 exhibited branching at both HFIP concentrations (Fig. 3B and C, arrows).
1883
3.3. Real-time observation of IAPP amyloid fibril growth To obtain further information, we performed real-time observations of fibril growth in the presence of 10% (v/v) HFIP at acidic pH (Fig. 4). Short IAPP fibrils were observed at 4 min and elongated at 20 min. Many fibrils crowded over the plate at 300 min. To compare the fibrils present at 150 min and 270 min, we merged the images of the two time periods colored green and red, respectively (Fig. 5). In this merged image, yellow fibrils were those that existed at 150 min, and red fibrils were newly elongated at 150–270 min. The merged image showed that new IAPP amyloid fibrils elongated from the ends of existing fibrils (Fig. 5-1), which suggested the presence of polarity in the growth direction. Interestingly, some red colors overlapped at the middle parts of preexisting fibrils (Fig. 5-2, and -3), suggesting that IAPP-Alexa532 not only interacted with growing ends, but also interacted with the lateral sides of preformed fibrils. Furthermore, branching of fibrils was detected (Fig. 5-3, and -4). These results also supported the hypothesis that IAPP monomers interact with the lateral sides of amyloid fibrils.
3.4. Observation with two-types of fluorescence dyes We used two types of fluorescence dyes with distinct wavelengths of fluorescence, i.e., ThT and Alexa532, to visualize the interactions between preformed fibrils and monomers (Fig. 6). Because non-labeled and Alexa-532-labeled IAPP formed similar AFM morphologies in the presence of HFIP (Fig. 2), it was difficult to distinguish them by AFM. On the other hand, TIRFM observation may provide images to address the mechanism of fibrillation at the single fibril level. The experiments were performed in 10 mM HCl and 10% (v/v) HFIP because both intact IAPP and Alexa532-labeled IAPP form amyloid fibrils easily. First, we prepared IAPP amyloid fibrils under the same solvent conditions and these fibrils were used as seeds. We then added IAPP-Alexa532 monomers to the IAPP seeds and observed the growth of amyloid fibrils. To precisely analyze the interactions, we changed the colors; green was for the fluorescence of ThT (originally blue) (Fig. 6A), and red was for the fluorescence of Alexa532 (originally orange) (Fig. 6B). We then merged these images (Fig. 6C). IAPPAlexa532 amyloid fibrils were expected to be yellow because green from ThT and red from Alexa532 were mixed. We first confirmed from the images obtained that IAPP-Alexa532 amyloid fibrils elongated from IAPP amyloid fibrils (seeds) (Fig. 6C, 1-3). Branching of IAPP-Alexa532 amyloid fibrils was observed
Fig. 2. Observation of IAPP fibrils by AFM. Twenty-five micromolar IAPP in 10 mM HCl and 10% (v/v) HFIP (A) or 25 μM IAPP-Alexa532 in 10 mM HCl and 30% (v/v) HFIP (B) were incubated for 1 day to make fibrils. Scale bars represent 1 μm.
1884
H. Yagi et al. / Biochimica et Biophysica Acta 1844 (2014) 1881–1888
Fig. 3. Amyloid fibrils of IAPP and IAPP-Alexa532 observed by TIRFM. A, fibrils of IAPP prepared by incubating for 1 day in 10 mM HCl and 10% (v/v) HFIP. B, and C, fibrils of IAPP-Alexa532 prepared by incubating for 1 day in 10 mM HCl and 10% (B) or 30% (C) (v/v) HFIP. The arrows show the branching of amyloid fibrils (see expanded images). Scale bars represent 10 μm.
(Fig. 6C-1). We also found that IAPP-Alexa532 amyloid fibrils elongated from both ends of the IAPP amyloid fibrils (Fig. 6C-2). Most importantly, the sides of preformed fibrils were mostly covered with Alexa532 fluorescence, exhibited by a yellow color. Without lateral binding, preformed fibrils should remain green in the merged images. We also show additional examples of overlapped images of Alexa532 and ThT, showing that a large number of preformed fibrils interacted with Alexa532-labeled IAPP (Fig. 6D). Therefore, these results provide strong evidence that IAPP-Alexa532 monomers interacted with the lateral sides of preformed fibrils as well as the elongating sites of amyloid fibrils. We expected IAPP-Alexa532 amyloid fibrils to be yellow, however, some fibrils remained red, which suggests that fluorescence resonance energy transfer (FRET) occurred between ThT and Alexa532. Consequently, ThT fluorescence became weak in the presence of IAPPAlexa532 monomers or fibrils, therefore, some fibrils exhibited a red color in merged images. Because FRET only occurs when the distance between the donor and acceptor molecules is very close, these results
are consistent with the hypothesis that IAPP fibrils and IAPP-Alexa532 monomers interacted with each other at the lateral sides of fibrils. 4. Discussion Surface diffusion at an interface between a solution and a protein crystal was previously shown to be an important elementary process of the crystal growth of substances [25]. Protein molecules form crystals in a supersaturated solution by nucleation followed by growth [40,41]. At above critical concentrations of solubility and in the presence of seed crystals, solute molecules started crystallization by adsorbing to the wide “terrace” of seed crystals (Fig. 7A-1). The bound molecules diffused randomly at the terrace, reaching the “step” and then “kink”, the growing sites of single crystals (Fig. 7A-2). The concentration of bound molecules was significantly higher than the bulk concentration, however, the diffusion constant decreased because of the strong interaction with the crystal surface. The gradual immobilization of bound molecules proceeded at the steps and kinks of crystals with further stabilizing
H. Yagi et al. / Biochimica et Biophysica Acta 1844 (2014) 1881–1888
1885
Fig. 4. Real-time observation of the elongation of IAPP amyloid fibrils. IAPP and ThT concentrations were 10 μM and 5 μM, respectively. Conditions were 10 mM HCl in the presence of 10% (v/v) HFIP at 25 °C. Scale bars represent 10 μm.
interactions, finally leading to the growth of crystals. This mechanism with surface diffusion has been linked to the overall rapid rate of crystallization in the presence of seed crystals in which surface diffusion is a critical step ensuring rapidity.
Considering the similarities between crystallization and the formation of amyloid fibrils [7], it is likely that a similar surface diffusion play important roles in the elongation process of amyloid fibrils (Fig. 7B-1). However, linear morphologies were often rigid and straight,
Fig. 5. Merged images of elongating fibrils at 150 min (green) and 270 min (red). Red amyloid fibrils are elongating fibrils. (1–4) are close-up views of the active sites of elongation. (1–2) shows elongation at both ends of the amyloid fibrils. (3–4) shows the branching of amyloid fibrils. The scale bar represents 10 μm.
1886
H. Yagi et al. / Biochimica et Biophysica Acta 1844 (2014) 1881–1888
Fig. 6. Two-color fluorescence observation of amyloid fibril growth. IAPP-Alexa532 monomers were added to IAPP amyloid fibrils prepared in 10 mM HCl and 10% (v/v) HFIP at 25 °C. A–C, fluorescence images detected by ThT (A), those by Alexa532 (B), and merged images (C). C1–3, IAPP-Alexa532 fibrils (red) elongated from IAPP fibrils (yellow). C1 and 3 show the branching of amyloid fibrils. C2 shows the elongation of IAPP-Alexa532 fibrils (red) from both ends of IAPP fibrils (yellow). (D) Additional examples of overlapped images of Alexa532 and ThT.
which suggested only the importance of the ends of fibrils for elongation (Fig. 7B-2). Real time imaging of fibril elongation also did not indicate the importance of regions other than the growing ends of fibrils. However, detailed analysis of the kinetics of fibril elongation indicated several pieces of evidence inconsistent with the simple growing mechanism, taking only the active ends of the fibrils into account. First, branching of fibrils from the trunk has been observed in several cases [17–20]. Regarding the seed-dependent growth of glucagon fibrils, branching from the preformed fibrils occurred frequently to form the brush-like morphology of fibrils [20]. Kinetic analysis with β2m amyloid fibrils at pH 2.5 monitored by tryptophan fluorescence indicated that the stoichiometry of “seed fibrils”:“bound monomers” was approximately 10:1 [21]. Considering the minimal size of seed fibrils made of several hundreds of monomers, this stoichiometry is too low to make binding sites at the ends of linear fibrils. The low stoichiometry can instead be easily explained by the lateral binding of monomers at the sides of fibrils. The lateral binding of β2m was further supported by an analysis of the intermediate of fibril elongation monitored by NMR [22,23]. We performed various NMR experiments to characterize the interaction of seed fibrils and acid-denatured monomeric β2m. An analysis of
transverse relaxation rates revealed that acid-denatured β2m underwent a structural exchange with an extensively unfolded form. The results suggested that the partially structured acid-denatured β2m was “activated” to become extensively unfolded, in which state the hydrophobic residues were exposed and associated with the seed. The low stoichiometry of seeds and bound monomers again supported the lateral binding model. We proposed a general model of amyloid elongation in comparison with the crystal growth of substances. Amyloid formation could only start under supersaturation above the solubility limit of amyloidogenic peptides. However, because of this supersaturation, it only started after a long lag time. Upon the addition of seed fibrils, fibril formation began immediately taking advantage of template-assisted reactions. As is the case with crystallization, the absorption of monomers occurred at the fibril surfaces. The major driving force of adsorption in amyloid fibrillation is hydrophobic interactions. However, these interactions were not strong enough to fix the bound molecules, therefore, they diffused on the fibril/water interface along with the dissociation/binding processes (Fig. 7B-1). When diffusing molecules reached the growing ends, tighter binding, coupled with conformational changes, occurred to finish one round of fibril growth (Fig. 7B-2). The model shown in
H. Yagi et al. / Biochimica et Biophysica Acta 1844 (2014) 1881–1888
1887
Fig. 7. Comparison of the surface diffusion-dependent formation of crystals and amyloid fibrils. A, schematic model of crystal growth. (1) Monomers interact with the terrace of crystals and undergo surface diffusion as well as dissociation/association. (2) They then bind to the step or kink to form crystals. (3) Alternatively, monomers associate at the terrace to form the island, a new nucleus for crystallization. B, schematic model of amyloid fibril growth. (1) IAPP monomers interact with the lateral sides of fibrils. Monomers repeat binding/dissociation and surface diffusion. (2) They then move to the growing ends to elongate fibrils. (3) Alternatively, monomers form the nucleus at the lateral side, and the branching of fibrils occurs.
Fig. 7B satisfactorily explains the mechanism of the branching of fibrils [17–20]. In the case of crystallization, several bound molecules on the same terrace diffused, associated with each other, and nucleated, leading to “island” growth (Fig. 7A-3). A similar association of monomers via surface diffusion may make a secondary nucleus, making branching possible (Fig. 7B-3). In amyloid fibrillation, the strong precipitation conditions were shown to lead to the formation of amorphous aggregates without specific interactions [24]. Amorphous aggregates correspond to glasses in crystallography. The formation of amorphous aggregates can also be explained on the basis of a surface diffusion model. Upon binding of monomers to seed fibrils, very strong interactions prevented the subsequent surface diffusion of bound monomers. Because binding continues to occur, this leads to the accumulation of trapped monomers and, consequently, the formation of amorphous aggregates. The same is true even in the absence of seed fibrils. Thus, lateral binding and surface diffusion are steps required for the formation of ordered amyloid fibrils, as is the case for the formation of single crystals. Finally, the mechanism of fibril elongation including the surface diffusion of laterally bound intermediates resembles the search for some DNA-binding proteins. Intramolecular as well as intermolecular translocation is considered to contribute significantly to the speed of a search for DNA-bound proteins [26–28]. Taken together, a sequential mechanism with weak binding, surface diffusion, and final tight docking may be a common mechanism maximizing the kinetic efficiency achieving highly organized interactions. By distinguishing the seed fibrils and newly formed fibrils of IAPP due to the introduction of two types of fluorescence probes, we succeeded in showing the lateral binding of monomers to preformed fibrils and the branching of fibrils from preformed fibrils. This type of lateral binding is well known to be an important elementary step for the crystallization of substances, i.e. surface diffusion. Surface diffusion determines the growth rate of crystals and their morphologies. Considering the similarities in the formation of amyloid fibrils and crystallization, it is clear that surface diffusion plays a role in the formation of amyloid fibrils, although the morphologies of fibrils
suggest the importance of the growing ends of fibrils. The inability of surface diffusion under very strong protein precipitation conditions leads to the formation of glass-like amorphous aggregates. Taken together, we proposed a new model of amyloid fibril growth in which lateral binding plays an important role. Further studies taking into account the roles of surface diffusion will advance our understanding of amyloid fibrils. Acknowledgements This work was supported by the Ministry of Education, Science, Sports, and Culture, Grant for Scientific Research, 22810014 (2010–2011), 24370067 (2012–2015), and 24592618 (2012–2015), Japan. We acknowledge financial support from Takeda Science Foundation.
References [1] F. Chiti, C.M. Dobson, Protein misfolding, functional amyloid, and human disease, Annu. Rev. Biochem. 75 (2006) 333–366. [2] J.D. Sipe, M.D. Benson, J.N. Buxbaum, S. Ikeda, G. Merlini, M.J. Saraiva, P. Westermark, Amyloid fibril protein nomenclature: 2012 recommendations from the Nomenclature Committee of the International Society of Amyloidosis, Amyloid 19 (2012) 167–170. [3] M.M. Barnhart, M.R. Chapman, Curli biogenesis and function, Annu. Rev. Microbiol. 60 (2006) 131–147. [4] V.A. Iconomidou, G. Vriend, S.J. Hamodrakas, Amyloids protect the silkmoth oocyte and embryo, FEBS Lett. 479 (2000) 141–145. [5] S.K. Maji, M.H. Perrin, M.R. Sawaya, S. Jessberger, K. Vadodaria, R.A. Rissman, P.S. Singru, K.P. Nilsson, R. Simon, D. Schubert, D. Eisenberg, J. Rivier, P. Sawchenko, W. Vale, R. Riek, Functional amyloids as natural storage of peptide hormones in pituitary secretory granules, Science 325 (2009) 328–332. [6] S.P. Graether, C.M. Slupsky, B.D. Sykes, Freezing of a fish antifreeze protein results in amyloid fibril formation, Biophys. J. 84 (2003) 552–557. [7] J.T. Jarrett, P.T. Lansbury Jr., Seeding “one-dimensional crystallization” of amyloid: a pathogenic mechanism in Alzheimer's disease and scrapie? Cell 73 (1993) 1055–1058. [8] H. Naiki, S. Hashimoto, H. Suzuki, K. Kimura, K. Nakakuki, F. Gejyo, Establishment of a kinetic model of dialysis-related amyloid fibril extension in vitro, Amyloid 4 (1997) 223–232.
1888
H. Yagi et al. / Biochimica et Biophysica Acta 1844 (2014) 1881–1888
[9] R. Wetzel, Kinetics and thermodynamics of amyloid fibril assembly, Acc. Chem. Res. 39 (2006) 671–679. [10] P.T. Lansbury Jr., B. Caughey, The chemistry of scrapie infection: implications of the ‘ice 9’ metaphor, Chem. Biol. 2 (1995) 1–5. [11] A.H. DePace, J.S. Weissman, Origins and kinetic consequences of diversity in Sup35 yeast prion fibers, Nat. Struct. Biol. 9 (2002) 389–396. [12] W.P. Esler, E.R. Stimson, J.M. Jennings, H.V. Vinters, J.R. Ghilardi, J.P. Lee, P.W. Mantyh, J.E. Maggio, Alzheimer's disease amyloid propagation by a templatedependent dock–lock mechanism, Biochemistry 39 (2000) 6288–6295. [13] P.H. Nguyen, M.S. Li, G. Stock, J.E. Straub, D. Thirumalai, Monomer adds to preformed structured oligomers of Aβ-peptides by a two-stage dock–lock mechanism, Proc. Natl. Acad. Sci. U. S. A. 104 (2007) 111–116. [14] T. Ban, M. Hoshino, S. Takahashi, D. Hamada, K. Hasegawa, H. Naiki, Y. Goto, Direct observation of Aβ amyloid fibril growth and inhibition, J. Mol. Biol. 344 (2004) 757–767. [15] T. Ban, K. Yamaguchi, Y. Goto, Direct observation of amyloid fibril growth, propagation, and adaptation, Acc. Chem. Res. 39 (2006) 663–670. [16] H. Yagi, T. Ban, K. Morigaki, H. Naiki, Y. Goto, Visualization and classification of amyloid β supramolecular assemblies, Biochemistry 46 (2007) 15009–15017. [17] S.B. Padrick, A.D. Miranker, Islet amyloid polypeptide: identification of long-range contacts and local order on the fibrillogenesis pathway, J. Mol. Biol. 308 (2001) 783–794. [18] F. Librizzi, C. Rischel, The kinetic behavior of insulin fibrillation is determined by heterogeneous nucleation pathways, Protein Sci. 14 (2005) 3129–3134. [19] M. Manno, E.F. Craparo, A. Podesta, D. Bulone, R. Carrotta, V. Martorana, G. Tiana, P.L. San Biagio, Kinetics of different processes in human insulin amyloid formation, J. Mol. Biol. 366 (2007) 258–274. [20] C.B. Andersen, H. Yagi, M. Manno, V. Martorana, T. Ban, G. Christiansen, D.E. Otzen, Y. Goto, C. Rischel, Branching in amyloid fibril growth, Biophys. J. 96 (2009) 1529–1536. [21] E. Chatani, R. Ohnishi, T. Konuma, K. Sakurai, H. Naiki, Y. Goto, Pre-steady-state kinetic analysis of the elongation of amyloid fibrils of β(2)-microglobulin with tryptophan mutagenesis, J. Mol. Biol. 400 (2010) 1057–1066. [22] T. Konuma, E. Chatani, M. Yagi, K. Sakurai, T. Ikegami, H. Naiki, Y. Goto, Kinetic intermediates of β(2)-microglobulin fibril elongation probed by pulse-labeling H/D exchange combined with NMR analysis, J. Mol. Biol. 405 (2011) 851–862. [23] K. Yanagi, K. Sakurai, Y. Yoshimura, T. Konuma, Y.H. Lee, K. Sugase, T. Ikegami, H. Naiki, Y. Goto, The monomer–seed interaction mechanism in the formation of the β2-microglobulin amyloid fibril clarified by solution NMR techniques, J. Mol. Biol. 422 (2012) 390–402. [24] Y. Yoshimura, Y. Lin, H. Yagi, Y.H. Lee, H. Kitayama, K. Sakurai, M. So, H. Ogi, H. Naiki, Y. Goto, Distinguishing crystal-like amyloid fibrils and glass-like amorphous aggregates from their kinetics of formation, Proc. Natl. Acad. Sci. U. S. A. 109 (2012) 14446–14451. [25] G. Sazaki, M. Okada, T. Matsui, T. Watanabe, H. Higuchi, K. Tsukamoto, K. Nakajima, Single-molecule visualization of diffusion at the solution — crystal interface, Cryst. Growth Des. 8 (2008) 2024–2031.
[26] J. Iwahara, G.M. Clore, Detecting transient intermediates in macromolecular binding by paramagnetic NMR, Nature 440 (2006) 1227–1230. [27] M. Doucleff, G.M. Clore, Global jumping and domain-specific intersegment transfer between DNA cognate sites of the multidomain transcription factor Oct-1, Proc. Natl. Acad. Sci. U. S. A. 105 (2008) 13871–13876. [28] P. Lian, L. Angela Liu, Y. Shi, Y. Bu, D. Wei, Tethered-hopping model for protein–DNA binding and unbinding based on Sox2–Oct1–Hoxb1 ternary complex simulations, Biophys. J. 98 (2010) 1285–1293. [29] P. Westermark, C. Wernstedt, E. Wilander, D.W. Hayden, T.D. O'Brien, K.H. Johnson, Amyloid fibrils in human insulinoma and islets of Langerhans of the diabetic cat are derived from a neuropeptide-like protein also present in normal islet cells, Proc. Natl. Acad. Sci. U. S. A. 84 (1987) 3881–3885. [30] G.J. Cooper, A.C. Willis, A. Clark, R.C. Turner, R.B. Sim, K.B. Reid, Purification and characterization of a peptide from amyloid-rich pancreases of type 2 diabetic patients, Proc. Natl. Acad. Sci. U. S. A. 84 (1987) 8628–8632. [31] J.D. Knight, J.A. Hebda, A.D. Miranker, Conserved and cooperative assembly of membrane-bound alpha-helical states of islet amyloid polypeptide, Biochemistry 45 (2006) 9496–9508. [32] S.M. Patil, S. Xu, S.R. Sheftic, A.T. Alexandrescu, Dynamic α-helix structure of micelle-bound human amylin, J. Biol. Chem. 284 (2009) 11982–11991. [33] A. Abedini, D.P. Raleigh, A critical assessment of the role of helical intermediates in amyloid formation by natively unfolded proteins and polypeptides, Protein Eng. Des. Sel. 22 (2009) 453–459. [34] R.P. Nanga, J.R. Brender, S. Vivekanandan, N. Popovych, A. Ramamoorthy, NMR structure in a membrane environment reveals putative amyloidogenic regions of the SEVI precursor peptide PAP(248–286), J. Am. Chem. Soc. 131 (2009) 17972–17979. [35] R.L. Hull, G.T. Westermark, P. Westermark, S.E. Kahn, Islet amyloid: a critical entity in the pathogenesis of type 2 diabetes, J. Clin. Endocrinol. Metab. 89 (2004) 3629–3643. [36] M.F. Engel, L. Khemtemourian, C.C. Kleijer, H.J. Meeldijk, J. Jacobs, A.J. Verkleij, B. de Kruijff, J.A. Killian, J.W. Hoppener, Membrane damage by human islet amyloid polypeptide through fibril growth at the membrane, Proc. Natl. Acad. Sci. U. S. A. 105 (2008) 6033–6038. [37] A. Quist, I. Doudevski, H. Lin, R. Azimova, D. Ng, B. Frangione, B. Kagan, J. Ghiso, R. Lal, Amyloid ion channels: a common structural link for protein-misfolding disease, Proc. Natl. Acad. Sci. U. S. A. 102 (2005) 10427–10432. [38] K. Yanagi, M. Ashizaki, H. Yagi, K. Sakurai, Y.H. Lee, Y. Goto, Hexafluoroisopropanol induces amyloid fibrils of islet amyloid polypeptide by enhancing both hydrophobic and electrostatic interactions, J. Biol. Chem. 286 (2011) 23959–23966. [39] T. Ban, D. Hamada, K. Hasegawa, H. Naiki, Y. Goto, Direct observation of amyloid fibril growth monitored by thioflavin T fluorescence, J. Biol. Chem. 278 (2003) 16462–16465. [40] Z. Kam, H.B. Shore, G. Feher, On the crystallization of proteins, J. Mol. Biol. 123 (1978) 539–555. [41] S.D. Durbin, G. Feher, Protein crystallization, Annu. Rev. Phys. Chem. 47 (1996) 171–204.
