Review Article (Invited)
Mechanisms of polyphosphate-induced amyloid fibril formation triggered by breakdown of supersaturation
2023 年 20 巻 1 号 論文ID: e200013
詳細
2023 年 20 巻 1 号 論文ID: e200013
Much effort has been devoted to elucidate mechanisms of amyloid fibril formation using various kinds of additives, such as salts, metals, detergents, and biopolymers. Here, we review the effects of additives with a focus on polyphosphate (polyP) on amyloid fibril formation of β2-microglobulin (β2m) and α-synuclein (αSyn). PolyP, consisting of up to 1,000 phosphoanhydride bond-linked phosphate monomers, is one of the most ancient, enigmatic, and negatively charged molecules in biology. Amyloid fibril formation of both β2m and αSyn could be accelerated by counter anion-binding and preferential hydration at relatively lower and higher concentrations of polyP, respectively, depending on the chain length of polyP. These bimodal concentration-dependent effects were also observed in salt- and heparin-induced amyloid fibril formation, indicating the generality of bimodal effects. We also address the effects of detergents, alcohols, and isoelectric point precipitation on amyloid fibril formation, in comparison with the effects of salts. Because polyP is present all around us, from cellular components to food additives, clarifying its effects and consequent biological roles will be important to further advance our understanding of amyloid fibrils. This review article is an extended version of the Japanese article, Linking Protein Folding to Amyloid Formation, published in SEIBUTSU BUTSURI Vol. 61, p. 358–365 (2021).
Polyphosphate (polyP), consisting of up to 1,000 phosphoanhydride bond-linked phosphate monomers, is one of the most ancient, enigmatic, and negatively charged molecules in biology. Here, we review the effects of additives with a focus on polyP on amyloid fibril formation of β2-microglobulin and α-synuclein, as well as the history and recent progress in polyP research in the field of molecular biology and protein science. Clarifying the effects and biological roles of polyP will be important for further understanding the mechanism of amyloid fibril formation.
Amyloid fibrils are ordered aggregates of misfolded protein associated with more than 35 human diseases, such as Alzheimer’s disease, Parkinson’s disease, and dialysis-related amyloidosis [1]. Amyloid fibrils are ~10 nm in diameter and several μm in length, with a characteristic cross-β structure [2,3]. Recent developments of cryo-EM revealed detailed structures of amyloid fibrils formed by β2-microglobulin (β2m) [4], α-synuclein (αSyn) [5], tau [6], amyloid-β 42 [7], and transthyretin [8]. The structures and topologies showed marked polymorphs, suggesting that they may be responsible for distinct pathogenesis even with the same amyloidogenic protein.
On the other hand, in the process of amyloid fibril formation, oligomeric intermediates, protofilaments, and liquid-liquid phase separation (LLPS) may play crucial roles in the pathology of amyloidosis [9–11]. These intermediate aggregates have been considered to be highly toxic and direct causes of diseases. Supersaturation is a kinetic condition in which the solution is trapped in an apparently soluble state, despite the solute concentration being greater than its thermodynamic solubility. Amyloid fibrils are generally formed by decreasing the solubility of precursor peptides and proteins by the addition of additives, combined with the disruption of supersaturation by agitation or seeding [12,13], and the concentration of solute decreases to its thermodynamic solubility. Further increases in driving forces of precipitation lead to rapid disruption of supersaturation and formation of glass-like amorphous aggregates [12,14]. To date, much effort has been devoted to elucidate the mechanism of protein aggregation in the presence of varying kinds of additives, such as salts (Figure 1), metals, detergents, and biopolymers.
Electroselectivity and Hofmeister series of anions. The strong anion-binding (left side) and weak anion-binding (right side) effects in the electroselectivity series. The salting-out and strong hydration effects (left side), where the native state is stabilized, and the salting-in and weak hydration effects (right side), where the denatured state is stabilized, in the Hofmeister series.
Polyphosphate (PolyP), consisting of up to 1,000 phosphoanhydride bond-linked phosphate monomers (Figure 2A), is one of the most ancient, enigmatic, and negatively charged molecules and is found in various microorganisms and human cells. In the history of polyP studies, Kornberg et al. developed a number of currently available methodologies to detect and purify polyP [15], and identified the enzyme responsible for bacterial polyP synthesis [16]. PolyP has been found to have a number of functions, e.g., an energy store and a reserve pool of inorganic phosphate mainly in bacteria [17], and the stimulation of blood coagulation [18], regulation of mitochondrial ion transport [19], energy metabolism [20], and the macrophage inflammatory response [21] in eukaryotes. Importantly, Cremers et al. [22] showed that polyP strongly promoted the amyloid fibril formation of several amyloidogenic proteins, and polyP has recently been focused on in research on protein aggregation [22–24]. Here, we reviewed the effects of polyP on amyloid fibril formation of β2m and αSyn under the ultrasonication. Although polyP is a negatively charged biopolymer, it is also a salt. We compared the effects of polyP, salts, and other additives such as detergents and alcohols on amyloid fibril formation, addressing their effects comprehensively.
Amyloid fibril formation of β2m in the presence of various types of polyPs. (A) Chemical structure of polyP; n is the degree of polymerization of the phosphate group. (B) Schematic models of the interaction between polyPs and β2m. At an acidic pH, the electrostatic interaction between polyPs and unfolded β2m (U) stabilized amyloid fibrils (AF) and amorphous aggregates (AA) competitively depending on the concentration of polyPs. At a neutral pH, under which β2m and polyPs were both negatively charged, hydration around the phosphate groups of polyP led to the preferential stabilization of dehydrated amyloid fibrils over amorphous aggregates. Although the native state was also likely to be stabilized, the net effects shifted the equilibrium toward the formation of amyloid fibrils. (C) Maximum values of thioflavin T (ThT) fluorescence (blue) and light scattering (LS) (red) at an acidic pH (top) and lag times of ThT fluorescence (blue) and LS (red) (bottom). Insets show TEM images of amyloid fibrils (Scale bars, 200 nm). (D) Phase diagrams for the polyP-induced aggregation of β2m. The aggregation of β2m depending on concentrations of polyP units (top) and inorganic phosphate (Pi) units (bottom) as a function of the chain length of polyP. The boundary lines between the monomer and amyloid and between amyloid and the amorphous aggregate are depicted in blue and red, respectively. The figure was modified from Zhang et al. [36].
More than 130 years ago, Hofmeister showed that the solubility of proteins increases or decreases depending on the types and concentrations of salts in aqueous solution [25,26]. It is known that these effects are more pronounced for anions than cations, since the effects of cations change depending on the sources. Three main mechanisms of the effects of salts on protein aggregation or stabilization have been investigated depending on the concentration of salts in detail [14,27]. The first is the Debye–Hückel ion screening effect, which is proportional to the square root of ionic strength, independent of salt species, and it generally plays a role at low ionic concentrations. The second is the counter ion-binding effect, which follows the electroselectivity series representing the affinities of charge–charge interactions at relatively lower concentrations of salts (Figure 1, top). This effect was examined based on the interaction between anions and resin in column chromatography [28,29]. The third is the Hofmeister effect, in which water structure-making kosmotropic ions (e.g., sulfate or phosphate anions) exert stronger effects than water structure-breaking chaotropic ions (e.g., thiocyanate or perchlorate anions) on the salting-out effect or protein stabilization at higher concentrations of salts (Figure 1, bottom) [25,30]. These three effects are distinguished from the concentration range and order of effectiveness of various salts [14,27].
β2m, which is a protein responsible for dialysis-related amyloidosis, forms amyloid fibrils at joints and the carpal tunnel in vivo [31]. To understand the mechanism of amyloid fibril formation of β2m, in vitro experiments have often been performed under acidic pH conditions, where β2m is denatured and readily converted into amyloid fibrils. Raman et al. [32] previously examined the effect of salts on amyloid fibril formation of β2m at pH 2.5. The effects of various anions followed the electroselectivity series, showing that amyloid fibril formation was induced by the counter anion-binding mechanism at a lower concentration of salts. Under neutral pH conditions, amyloid fibril formation did not readily occur, because of the native structure of β2m. It has been reported that several β2m mutants, such as D76N [33] and V27M [34] mutants, form amyloid fibrils at a neutral pH and that heating of wild-type β2m solution up to 60°C facilitates amyloid fibril formation at a neutral pH. These results indicated that destabilization of the native structure is linked with amyloid fibril formation [13,35], although the detailed mechanism at a neutral pH remains unclear.
Amyloid fibril formation of β2m was investigated using varying chain lengths of polyP at acidic and neutral pHs [36]. Amyloid fibril formation at an acidic pH was markedly accelerated by the addition of polyP, and the optimal polyP concentration for amyloid fibril formation decreased with an increase in the chain length of polyP (Figure 2C, D). PolyP and β2m have negative and positive charges, respectively, at an acidic pH, and the electrostatic interactions between them shield the charge repulsion and induce amyloid fibril formation (Figure 2B). Any further increase in the driving force for aggregation induced amorphous aggregates at higher concentrations of polyP. Under neutral pH conditions, long polyP with 60 to 70 phosphates significantly induced the amyloid fibril formation of β2m at several μM, a similar concentration range to that in vivo. Both polyP and β2m have negative charges at a neutral pH, and polyP appears to preferentially hydrate around β2m. Because preferential hydration is energetically unfavorable and the unfolded conformations have a larger exposed surface than the folded native conformation, the compacted folded native conformation can be stabilized. When the protein concentration is high enough to make intermolecular interactions possible, preferential hydration stabilizes less water-exposed amyloid fibrils (Figure 2B), as will be described later.
The affinity between polyP and β2m was directly confirmed using isothermal titration calorimetry experiments. After the injection of β2m into polyP solution, an exothermic reaction which was dominantly caused by strong electrostatic binding was observed at an acidic pH. On the other hand, an endothermic reaction, which could be caused by entropy-driven reaction through the dehydration of water molecules around β2m, were observed at a neutral pH. Thus, the mechanisms of polyP-induced amyloid fibril formation of β2m change depending on the pH of the solution.
αSyn, which is an intrinsically disordered protein consisting of 140 amino acid residues, is a causative protein for synucleopathies including Parkinson’s disease, multiple system atrophy and dementia with Lewy bodies [37]. Munishkina et al. [38] previously showed that amyloid fibril formation of αSyn was accelerated by the addition of varying kinds of salts at a neutral pH. These effects were followed by the Hofmeister series at relatively higher concentrations of salts and the anion-binding mechanism at lower concentrations of salts, although the acceleration of amyloid fibril formation also depended on the kinds of cation [38].
Various physicochemical approaches were employed to investigate the effects of polyP using varying chain lengths on αSyn under ultrasonication [39]. The orthophosphate and diphosphate showed a single optimal concentration of amyloid fibril formation, although the triphosphate and longer-chain phosphates showed two optima, with the second at a concentration lower than that of orthophosphate or diphosphate (Figure 3A). The optima at lower polyP concentrations could be caused by counter anion-binding between negatively charged phosphate groups and positive charges of αSyn, and the optima at higher polyP concentrations could be caused by preferential hydration in the Hofmeister salting-out effects of phosphate groups (Figure 3B). Thus, polyP induced amyloid fibril formation of αSyn by distinct concentration-dependent mechanisms.
Amyloid fibril formation of αSyn in the presence of polyPs. (A) Amyloid fibril formation in the presence of varying chain lengths and concentrations of polyPs plotted against polyP or NaCl concentrations. (B) Logarithmic plots of optimal amyloid concentrations, which are caused by the charge–charge interactions (orange) and Hofmeister salting-out effects (green), obtained using the polyP concentration (closed circles), phosphate unit concentration (open circles), and NaCl concentration (closed triangles). (C) Schematic of the mechanisms for polyP-induced amyloid fibril formation of αSyn at a neutral pH. At lower concentrations of tetraP, shielding of the charge repulsions leads to compact NAC regions stabilized by hydrophobic interactions and the formation of low-salt amyloid fibrils. At higher concentrations of tetraP, hydrated tetraPs cause Hofmeister salting-out effects and lead to the formation of high salt amyloid fibrils. The figure was modified from Yamaguchi et al. [39].
NMR titration experiments showed that negatively charged tetraphosphates interacted with positively charged “KTK” segments in four KTKEGV repeats of αSyn at lower concentrations of tetraphosphate (Figure 3C). At higher concentrations, hydrated tetraphosphates affected the surface-exposed hydrophilic groups of compact αSyn. There are several reports on the role of KTKEGV repeats in αSyn [40], possibly interacting with negatively charged biomolecules in vivo. Structural analysis using cryo-EM suggests that phosphate ions are located at around K43, K45, and K58 of αSyn [5]. αSyn may interact with negatively charged molecules such as polyP, triggering amyloid fibril formation in vivo.
Amyloid fibril formation of β2m was markedly accelerated by the addition of polyP via counter anion-binding and preferential hydration mechanisms at lower and higher polyP concentrations, respectively, in distinct concentration-dependent manners. These bimodal concentration-dependent effects were also observed in salt- [38,41] and heparin-induced amyloid fibril formation [42,43].
Anion binding-dependent amyloid fibril formation was also revealed by studying the effects of strong acids (e.g., HCl or H2SO4) and their salts (e.g., NaCl or Na2SO4) on β2m under ultrasonication [44]. Treatments with these strong acids promoted amyloid fibril formation through the anion-binding mechanism because similar promotion was observed at pH 2 by the addition of Na salts at the same concentrations. In the folding at an acidic pH, the formation of a molten globule state, a compact intermediate with native-like secondary structures, was induced by the shielding of positive charge repulsion with anions [44], as previously reported with other proteins [27,45]. Thus, the anion-binding mechanism induces the formation of the molten globule state and amyloid fibrils by intra- and intermolecular reactions, respectively, upon decreasing the solubility of unfolded proteins.
Meanwhile, the preferential hydration in the Hofmeister series enables the stabilization as well as salting-out of proteins simultaneously [46–50]. To address these two effects at high concentrations of salt (~1 M), protein denaturation is considered using a two-state model (Mechanism 1):
| (Mechanism 1) |
| (1) |
The equilibrium constant (KD) (Eq. (1)) is a function of the concentration of the additive (S). The effect of the additive on the equilibrium constant can be shown by the difference between additives bound to denatured and native proteins in Eq. (2):
| (2) |
where aS is the activity of the additive S, and ΔνS is the difference between additive bound to the denatured (νD) and native (νN) states of protein. In the stabilization of protein, the equilibrium constant KD decreases with an increase in the concentration of the additive, i.e., ΔνS is negative in Eq. (2), indicating that additives bind to the native state compared with the denatured state.
Eq. (2) is extended to include the contribution of water. The equilibrium binding experiment does not measure the total amount of additive bound to the protein, but the relative affinities of additive and water for the protein. There is an obligatory relationship between the changes in activity of the additive and water:
| (3) |
where mS is the molarity of additive S, and ΔνW is the difference between the number of water molecules bound to the denatured and native forms of the protein, respectively. Eq. (3) reduced to Eq. (2) at relatively low concentrations of S, i.e., when mS <<55.5. The preferential binding of additive S to a protein (ΔνS,pref) is the amount of S bound to the protein in excess of S and solvent that may be bound in the same proportion in which S and solvent are present in the solvent mixtures. Preferential binding can be positive or negative, with negative values representing excess hydration (Figure 4A).
Schematic representation of preferential hydration. (A) Distribution of water molecules (open circles) and additives (closed circles). For the preferential hydration, the additive concentration in hydrated shell is lower than that in the bulk. (B) Less water-exposed native state or amyloid fibrils are stabilized in the presence of a stabilizing additive (e.g. kosmotropic anion). The area of stabilizer-induced preferential exclusion (gray region) becomes greater as the protein-solvent interface increases during denaturation. The figure was created on the basis of the basic model in references [50,51].
The dialysis equilibrium technique has been employed to measure the amount of additive bound to proteins in the native state [51], and binding (ν) is defined as:
| (4) |
where [Lin] and [Lout] are the concentrations of ligand inside and outside the dialysis membrane, respectively. All the binding stoichiometries for the stabilizer (e.g., kosmotropic anion) are negative, showing that the ligand concentration inside the dialysis membrane is lower than that outside the membrane. In contrast, the binding stoichiometries for the destabilizer (e.g., chaotropic anion) are positive. In the protein denaturation, when the interactions are defined by the surface of contact with the solvent, the area of stabilizer-induced preferential exclusion increases upon denaturation (Figure 4B) [46–51]. This is thermodynamically unfavorable, i.e., the addition of stabilizer increases the chemical potential of the protein, which leads to the preferential stabilization of less water-exposed native state or amyloid fibrils. Meanwhile, the area of preferential binding induced by the destabilizer increases with a protein denaturation, which is thermodynamically favorable, leading to conversion into a more denatured state.
We also address the effects of other additives such as detergents [52–54] and alcohols [55,56] on amyloid fibril formation (Figure 5). Sodium dodecyl sulfate (SDS), which is one of anionic detergents, promoted amyloid fibril formation of β2m [53] and αSyn [52,54] at slightly below the critical micelle concentration (CMC) under neutral pH conditions (Figure 5C). At higher concentrations than CMC, both β2m and αSyn showed α-helical conformation by the formation of intermolecular hydrogen bonds under the hydrophobic environment of the micelle interior. The helical structures are not compactly collapsed, but extended by interacting with the hydrophobic SDS micelle interior. Meanwhile, moderate concentrations of alcohols such as 2,2,2-trifluoroethanol (TFE) and 1,1,1,3,3,3-hexafluoroisopropanol (HFIP) promote amyloid fibril formation of β2m [55] and its fragmented peptide K3, corresponding to Ser20 to Lys41 of β2m (Figure 5C) [56]. These alcohols are known to form dynamic clusters in aqueous solution at approximately 30% (V/V) [57]. TFE and HFIP promoted amyloid fibril formation at these concentrations by strengthening hydrophobic interactions [56,58]. At higher alcohol concentrations, proteins lead to extended α-helical conformation by forming intramolecular hydrogen bonds in a manner similar to SDS.
Distinct mechanisms of amyloid fibril formation. (A) Counter ion-binding mechanism observed under acidic conditions in the presence of low concentrations of salts. (B) Salting-out mechanism observed under high salt conditions independent of pH. (C) Hydrophobic additive-binding mechanism observed in the presence of moderate concentrations of alcohols or detergents like SDS. (D) pI-precipitation mechanism. To form amyloid fibrils, van der Waals interactions are also required. Common to these conditions, amyloid fibrils form above solubility of precursor proteins upon breaking the supersaturation. The figure was reproduced from Furukawa et al. [59].
In addition, amyloid fibril formation of αSyn was accelerated at its isoelectric point (pI) of 4.7 in the absence of salt [59]. This was caused by intermolecular attractive charge-charge interactions, where αSyn has ±17 charges even with a zero net charge (Figure 5D). The intermolecular attractions increase the size of the molecules, leading to decrease in solubility.
In general, a polypeptide in an aqueous environment collapses in the early stages of protein folding because of decreased solubility, leading to coil-globule transition [60]. The coil-globule transition leads to a compactly folded intermediate and finally a native state stabilized by intramolecular hydrophobic and hydrogen bond interactions and van der Waals interactions. In the presence of anions, both the counter anion-binding and preferential hydration contribute to strengthening these interactions at lower and higher anion concentrations, respectively [36,43]. On the other hand, both detergents and alcohols at concentrations slightly below CMC and clustering concentrations, respectively, strengthen hydrophobic and hydrogen bond interactions before complete dissolution in hydrophobic environments [54–56]. At around pI, the protein tends to show compact conformation through intramolecular interaction [59]. These various conditions decrease protein solubility. When combined with van der Waals interactions, a protein folds to a native state or misfolds to amyloid fibrils depending on the balance of intra- and intermolecular interactions, and amyloid fibrils are formed after breakdown of the supersaturation.
In addition, to understand the additive-dependent promotion and inhibition mechanisms in amyloid fibril formation, we have applied two classical approaches employed by Tanford and colleagues, which have been used traditionally to explain the additive-dependent conformational transitions of proteins, i.e., (i) a ligand binding-dependent mechanism and (ii) cosolute-dependent mechanism (Figure 6) [54]. The cosolute-dependent mechanism assumes that the solubility of proteins changes depending on the cosolute concentration. Although two types of models are exchangeable, either of the two is chosen depending on the size of ligands or affinity of the interactions. These models may explain the amyloid fibril formation of αSyn on the lipid membrane, and more complicated cases, e.g., NaCl- or heparin-dependent amyloid fibril formation, where various additives accelerate and then inhibit amyloid fibril formation in a concentration-dependent manner [61]. Comprehensive understanding based on the solubility- and supersaturation-dependent mechanisms would be necessary to clarify amyloid fibril formation.
Alternative models for additive-dependent amyloid fibril formation. Illustration of (i) the ligand binding- and (ii) cosolute-dependent solubility-modulation model to simulate additive-dependent amyloid fibril formation. The figure was reproduced from Sawada et al. [54] with permission. Copyright 2020 American Chemical Society.
Recently, studies on polyP have advanced markedly in the field of molecular biology and protein science. Wang et al. [62] showed that polyP interacts with positively charged green fluorescence protein, resulting in liquid-liquid phase separation (LLPS) by intermolecular electrostatic interactions inside cells. Beaufay et al. [63] revealed that polyP serves as a new driver for heterochromatin formation in bacteria. PolyP, DNA, and Hfq, which is the nucleoid-associated protein, form three-component liquid droplets. In contrast, Azevedo et al. [64] reported that inorganic polyP can be non-enzymatically and covalently attached to the yeast protein Nsr1 and its interacting partner Top1. This “polyphosphorylation” negatively regulates their interaction and impairs Top1 enzymatic activity. Thus, the negatively charged polyPs may play an important role in vivo and associate with LLPS as well as protein aggregation, although positively charged biopolymers have also promoted amyloid fibril formation of αSyn in vivo [65].
Meanwhile, polyP has often been used as a food moisturizing agent, applied to rice balls, ham, wieners, and other foods. PolyP consists of numerous phosphate groups, water-structures forming kosmotropic anions, retaining water molecules around them following application on foods. In protein aggregation, polyP is a useful molecule which essentially stabilizes the native structure of protein, but it can simultaneously induce amyloid fibril formation by decreasing the solubility combined with the breakdown of supersaturation. Thus, native conformation and amyloid fibrils may be two sides of the same coin, and comprehensive understanding based on the solubility- and supersaturation-dependent mechanisms will be essential.
Here, we focused on the effects of polyP and other additives such as heparin, detergents, and alcohols and the effect of pI precipitation on amyloid fibril formation. These additives or pI conditions induce unfolded protein conversion into a more compact structure by decreasing solubility, thus stabilizing the native state. At a high protein concentration, the same conditions alternatively lead to amyloid fibril formation by the intermolecular interactions coupled with the breakdown of supersaturation. Ultrasonication is one of the most powerful agitations to break the persistent metastability of supersaturation [66–70]. So far, we have developed a Handai amyloid burst inducer (HANABI) by combining a water bath-type ultrasonicator and microplate reader, which has made it possible to automate amyloid fibril formation of amyloidogenic proteins [68]. Moreover, controlling the oscillation amplitude and frequency of each transducer, we have developed an optimized sonoreactor, which has improved reproducibility and the control of amyloid fibril formation (Figure 7) [69–71]. Although the protein-misfolding cyclic amplification (PMCA) makes it possible to detect pathological amyloid fibrils formed in vivo with high sensitivity using ultrasonication [72,73], the HANABI system, in addition, effectively stimulates the solution to disrupt the supersaturation state and promote spontaneous amyloid fibril formation. This enables further early-stage diagnosis even without seeds through accelerated primary nucleation (i.e., identification of susceptibility risk biomarkers [74]) and a comprehensive understanding of supersaturation-dependent amyloid fibril formation [71].
Ultrasonication-forced amyloid fibril formation inducer, HANABI-2000 system. (A) 3D schematic illustration of the multi-channel optimized sonoreactor for amyloid-fibril assay, HANABI-2000. The dimensions of the device are 500×550×550 mm3. (B) Block chart of the control units of the HANABI-2000. The figure was reproduced from Nakajima et al. [69].
The authors declare that they have no conflicts of interest regarding the contents of this article.
All authors discussed and wrote the manuscript.
The evidence data generated and/or analyzed during the current study are available from the corresponding author on reasonable request.
This study was performed as part of the Cooperative Research Program for the Institute for Protein Research, Osaka University (CR-21-02), and was supported by the Japan Society for the Promotion of Science (20K06580 to K.Y., 21K19224 and 22H02584 to Y.G., 22K14013 to K.N., and Core-to-Core Program (JPJSCCA20180007) to Y.G.), and JKA and its promotion funds from AUTORACE to K.Y. and K.N.
