Volume 52, Issue 4 p. 281-292
Review
Free Access

Polymer physics inspired approaches for the study of the mechanical properties of amyloid fibrils

Lisa R. Volpatti,

Department of Chemistry, University of Cambridge, Lensfield Road, CB2 1EW, United Kingdom

Search for more papers by this author
Tuomas P. J. Knowles,

Corresponding Author

Department of Chemistry, University of Cambridge, Lensfield Road, CB2 1EW, United Kingdom

Correspondence to: T. P. J. Knowles (E-mail: tpjk2@cam.ac.uk)Search for more papers by this author
First published: 07 December 2013
Citations: 11

ABSTRACT

Amyloid structures constitute a class of highly ordered nanomaterials formed by insoluble protein aggregates. These aggregates are characterized by a cross-β structural motif in which β-sheets are oriented perpendicular to the fibril axis and bound together by a dense hydrogen bonding network. Although they have been associated with several neurodegenerative disorders, such as Alzheimer's and Parkinson's diseases, amyloid fibrils have also been found in many physiologically beneficial roles, for instance in adhesives and hormone storage. Inspired by this natural occurrence of functional amyloid, the hierarchal self-assembly of these structures has recently been used to develop artificial biomaterials for applications in medicine and nanotechnology. In order to realize the full potential of amyloids as functional materials, it is important to understand their fundamental mechanical properties. This review explores a range of experimental strategies to determine the mechanical properties of amyloid fibrils and discusses the results in the context of polymer physics concepts. © 2013 Wiley Periodicals, Inc. J. Polym. Sci., Part B: Polym. Phys. 2014, 52, 281–292

INTRODUCTION

A diverse array of proteins has been shown to possess the propensity to convert from their soluble native states to form insoluble linear aggregates, amyloid fibrils. Amyloids were first discovered in association with neurodegenerative diseases, such as Alzheimer's, Parkinson's, Huntington's, and Creutzfeldt–Jakob diseases.1-12 Although they have been largely associated with pathogenicity, amyloid structures have also been found in many physiologically functional roles,13 such as adhesives,14-17 bacterial coatings,18 and storage for peptide hormones.19 The discovery of functional amyloid fibrils in nature has provided the inspiration to explore possibilities for their use as the basis for artificial protein-based functional materials.20 Thus, measuring and quantifying the mechanical properties of amyloid fibrils allows a materials science based understanding to be developed of an important class of materials that are involved in both biological function and malfunction and hence may provide insights into their role in diseases in addition to demonstrating their potential for use as bionanomaterials in various technological applications.

Amyloid structures represent a class of materials with a wide range of rigidities that are characterized by a cross-β structural motif in which β-sheets are oriented perpendicular to the fibril axis.21-23 While this characteristic cross-β structure was discovered over half a century ago,24 detailed structural determination has only recently become possible,22 and studies within the past decade have brought to light further structural detail including the existence of a steric zipper, which is formed by side chain interactions in the double β-sheet core of the fibril spine.25, 26 The constituent β-strands self-assemble into a hierarchy of oligomers, protofilaments, and mature fibrils. These latter species are typically a few nanometers in diameter and micrometers in length but aggregate to form ordered supramolecular structures such as amyloid biofilms or plaques.

From a structural point of view, therefore, amyloid fibrils represent a supramolecular polymer formed from protein units assembled into linear structures through non-covalent interactions between the subunits. Consequently, the approaches for describing the statistical mechanical properties of synthetic polymers27, 28 that have proven to be highly successful in the context of covalent biopolymers such as DNA29, 30 are likely to have great potential to elucidate the nature of the material properties of amyloid fibrils. In this review, we consider amyloid fibrils and other biological polymers in the light of polymer physics to provide a summary of their nanomechanical properties and an overview of the utilization of these properties in natural and artificial materials.

BEHAVIOR OF SEMIFLEXIBLE POLYMERS

A polymer molecule in solution exhibits Brownian motion; it fluctuates constantly as a result of interactions between the polymer and thermally excited solvent molecules.31, 32 These fluctuations depend on the fundamental physical properties of the polymer such as its bending rigidity, and hence the study of the statistical mechanics of such fluctuations is a valuable route towards understanding the nanoscale mechanics of polymers and other linear structures. The coarse-grained worm-like chain (WLC) model is commonly used to describe semi-flexible polymers of a biological nature.33-37 In this model, the polymer is characterized as a continuous chain whose elastic behavior is governed by its bending rigidity, or the measure of the polymer's resistance to bending forces.

As depicted in Figure 1, the persistence length, lp, is defined as the arc length, s, along the polymer chain at which the tangential angles Θ(0) and Θ(s) become uncorrelated.38 Analogously, it can be thought of as the average length that the polymer chain changes direction as a result of random thermal fluctuations.39 The persistence length is inversely proportional to the bending rigidity, lp = κ/kBT. For isotropic elastic rods, the bending rigidity can be readily deconvoluted into an elastic (Young's) modulus E and a geometric factor, the cross sectional moment of inertia I, κ = EI.

image
A freely fluctuating polymer of contour length L. At arc length s, t(s) represents the tangent vector, Θ(s) is the corresponding tangent angle. The arc length at which Θ(0) and Θ(s) are no longer correlated represents the persistence length, lp.

Within this framework, an important issue that arises in the study of amyloid fibrils is the challenge of measuring the cross sectional area which is required to relate the measured persistence length to the Young's modulus. In some cases, where the fibril structure is known, the cross-sectional area can be directly measured.40-42 In many instances, however, this is not possible, and it becomes necessary to propose a filament packing mechanism. For closely packed filaments, the cross-sectional area can be calculated by approximating the protofilament geometry as a cylinder of radius r that can be estimated from atomic force microscopy (AFM) images.42, 43 For a ribbon-like packing scheme of multistranded fibrils, the cross section can be approximated by a rectangle of dimensions 2r by 2nr, where r is the radius of a single filament and n is the number of filaments comprising the fibril.43 Obtaining the radius by measuring the width of fibrils presents additional inaccuracies due to sample convolution effects. Therefore, the height is a much more reliable parameter that can be readily extracted from AFM images since it is not dependent on the tip geometry.44

Larger sample cross sectional areas result in larger values of I and subsequently longer persistence lengths.39 Polymers and filaments can broadly be characterized as flexible or stiff by comparing their persistence length to their contour length, L, which is the largest possible value of the end-to-end distance for a fully stretched out chain.33 Polymers with persistence lengths much greater than the contour length, lpL will appear very rigid, while those with persistence lengths much smaller than the contour length lpL will appear very flexible. Thus, longer persistence lengths correlate to stiffer structures. When the persistence length is on the same order of magnitude as the contour length, lp ∼ L, the polymer is referred to as semiflexible.45

Measuring Mechanical Properties of Single Molecules and Filaments

The bending rigidity and persistence length can be determined experimentally by statistically analyzing the shape fluctuations of a freely fluctuating semiflexible polymer.42 Measuring the end-to-end distance and fluctuations of a polymer confined in a channel can also provide insights on its persistence and contour lengths.46

Over the past several decades, single-molecule techniques have developed into powerful tools to probe the mechanical properties of polymers. Notably, single-molecule experiments of biological polymers have gained popularity due to their relevance to diseases and potential applications in biotechnology. While the application of polymer physics to amyloid systems has been the topic of much research in the past few years, this general approach has historically been highly successful in elucidating the fundamental nanomechanics of other biomolecules, including DNA and sickle hemoglobin (HbS), which are discussed briefly below.

Biomechanics of DNA

As a semiflexible polymer, double stranded (ds)DNA is governed by the WLC theory in the entropic elasticity regime consisting of mechanical loads up to around 35 pN.47, 48 Thus, the WLC theory is often used to describe the physics of DNA in many physiological roles although there are certain applications in which the behavior of DNA deviates from this model. For instance, transverse fluctuations of single λ-phage DNA molecules observed by fluorescence microscopy suggest that the force-fluctuation relation is described by the WLC theory in the strong force regime; however, in low force regimes it was determined that the experimental data more closely follows values for an ideal ring polymer.49 Additionally, the mechanical properties of DNA molecules on short length scales deviate significantly from those predicted by the WLC model as revealed by AFM,50-52 single-molecule cyclization,53 and probing geometric fluctuations at various length scales using trajectories from molecular dynamics simulations.54

Direct measurement of shape fluctuations using a small-angle X-ray scattering (SAXS) interferometry technique yielded a bending persistence length of 55 ± 10 nm for dsDNA under physiological conditions.55 This measurement agrees with the widely accepted value of ∼50 nm that has been previously obtained by various methods,47, 56 suggesting that SAXS can be used to determine the elastic characteristics of freely diffusing polymers. In addition to direct observations, a principal-components analysis can be applied to study near-equilibrium shape fluctuations of single DNA molecules. The results of one such analysis revealed that dynamic internal interactions that depend on the relative motion of chain elements were prevalent over static interactions that depend on the relative position of the elements.57

A novel approach to measuring the nanoscale mechanical properties of macromolecules in situ was recently reported using 4D electron microscopy.58 In this method, a laser pulse induces substrate vibrations which in turn mechanically excite the vibrational oscillations of a DNA structural network. By observing these oscillations in space and time, information about the specimen's nanomechanical properties can be acquired. With this method, the Young's modulus of λ-DNA nanostructures was found to be 15 ± 3 GPa, which is much greater than the previously reported value of around 300 MPa56, 59 that was determined from single-molecule stretching experiments in solution, suggesting that the nanostructures were dehydrated.58 Using the aforementioned techniques to understand the nanomechanics of individual molecules of DNA is especially important in the field of DNA nanotechnology, as DNA nanostructures are being incorporated in biomimetic systems, diagnostics, therapeutics, and even photonics.60, 61

Sickle Hemoglobin Mechanics

Under deoxygenated conditions, HbS molecules polymerize to form rigid fibers that damage the membrane and cytoskeleton of red blood cells (RBCs), causing them to deform and adapt the conformation characteristic of sickle cell disease.62-65 Using differential interference contrast microscopy to observe thermal fluctuations in fiber shape, the persistence lengths were determined to vary from 0.24 mm for a single fiber to 13 mm for fiber bundles.66 While a single fiber the size of a RBC has a Young's modulus of about 100 MPa and would buckle under 0.175 pN66 of force, bundles with persistence lengths 1600 times the diameter of a RBC could easily deform the membrane.65 The same study showed that the Young's modulus for gels with a HbS concentration of 24.4 g/dL was 0.10 GPa,65 while a novel microrheological analysis of gels with HbS concentrations ranging from 6 to 12 g/dL yielded moduli of 300 to 1500 kPa, respectively.67 By measuring bending fluctuations and twist fluctuations with electron microscopy, the bending persistence length of a single fiber was determined to be 0.13 mm (on the same order of magnitude as the previously reported value of 0.24 mm), and the torsional persistence length was found to be a mere 2.5 μm, suggesting anisotropy in HbS fibers.68 Using this information of HbS nanomechanics, several coarse-grained models have been proposed to simulate the mechanical properties of HbS fibers and predict the resultant RBC morphology.69-72

NANOMECHANICS OF AMYLOID FIBRILS

Polymer physics inspired concepts have been highly successful in quantifying the mechanical properties of covalent biopolymers such as DNA and noncovalent assemblies such as HbS as outlined above. The focus of this article is to review the emerging use of such approaches to shed light on the nanoscale properties of amyloid fibrils. The mechanical properties of single fibrils have largely been observed with AFM using techniques such as nanoindentation, quantitative nanomechanical mapping, force spectroscopy, and the statistical analysis of shape fluctuations.

Nanoindentation

Nanoindentation, the method most universally used to obtain the nanomechanical characteristics of a specimen, involves subjecting the polymer to a contact force from an AFM tip to obtain a force versus separation curve.44 In conjunction with mechanical modeling (such as the Hertz model) the force-separation curve can be used to extrapolate mechanical properties such as Young's modulus and hardness.73, 74 A nanoindentation study on the elastin-like polypeptide poly(ValGlyGlyLeuGly) showed that the force vs. distance curves were equivalent irrespective of the tip position along the fiber axis, suggesting that the fibers preferably buckle rather than deform by bending.75 This finding supports the previous conclusion that polypeptides of the form poly(XxxGlyGlyZzzGly) (Xxx, Zzz = Val, Leu) self-assemble into stiff amyloid-like fibers.76 Further confirmation that poly(ValGlyGlyLeuGly) adopts a β-sheet structure arises from another AFM study that estimates the fibrils' Young's modulus to range from 3.5 to 7 MPa, much higher than that of elastin which is around 1 MPa.77 This large range in moduli is common for amyloid fibrils and indicative of their structural heterogeneity. For example, AFM-based nanoindentation studies estimated that the elastic moduli of glucagon amyloid fibrils range from 0.72 ± 0.80 GPa to 1.26 ± 0.62 GPa under small compressive forces78 while the moduli of insulin amyloid fibrils ranged from 5 to 50 MPa.79

Peak Force Quantitative Nanomechanical Mapping (PF-QNM)

A relatively new AFM technique similar to nanoindentation, PF-QNM can be used to directly measure surface properties of nanomaterials with high spatial resolution and surface sensitivity.56 Figure 2 shows that the penetration depth of the AFM tip in this method is an order of magnitude smaller than that needed for nanoindentation. As opposed to simple tapping mode AFM, where the feedback parameter is dependent solely on the amplitude of the tip relative to the sample, PF-QNM uses the peak force value of the force-distance curve as the parameter in its feedback loop allowing for mechanical properties to be measured in real time.44 For instance, the modulus of elasticity can be readily mapped and exported according to the Derjaguin-Muller-Toporov (DMT) model.80, 81

image
AFM-based methods of measuring the nanomechanical properties of individual fibrils. (a) The cantilever is displaced vertically until a predetermined force is met to produce a force versus separation curve. A mathematical model is subsequently fitted to the curve to determine certain mechanical properties of the fibril. (b) Similar to nanoindentaiton, the AFM tip is displaced vertically towards the fibril. However, the penetration depth in PF-QNM is around 2 nm, much smaller than nanoindentaiton distances on the order of tens of nanometers. The force versus separation curves are often fitted with a DMT model to determine the fibril's mechanical properties. (c) AFM-based force spectroscopy involves the adsorption of the fibril to the AFM tip upon contact to yield a force versus extension curve. This curve is then fitted with an appropriate model to obtain the mechanical properties of the fibrils. (d) Topographic data is produced by repeatedly passing the AFM tip in tapping mode over the fibril. Statistical analysis of these AFM images yields the mechanical properties of the fibril. Reproduced from Ref. 42, with permission from AAAS.

Using PF-QNM Adamcik et. al. found the Young's modulus of β-lactoglobulin amyloid fibrils to be 4.98 GPa,82 which is consistent with the value previously reported by statistical analysis of thermal fluctuations.42 Values of Young's moduli for different classes of amyloid fibrils including α-synuclein, Aβ (1–42), bovine serum albumin, insulin, lysozyme, ovalbumin, and Tau protein all fall in the range of 2–4 GPa.83-85 These findings suggest that PF-QNM provides an accurate and reliable method of directly measuring the nanomechanical properties of single amyloid fibers that provides immediate results without the statistical analysis of a large number of AFM images.

Force Spectroscopy

In AFM-based force spectroscopy, the tip adsorbs the molecule and stretches it upon retraction producing a resultant force-extension curve that can be fitted to a theoretical model, such as the WLC model, to determine the molecule's mechanical properties [Fig. 2(c)].86-89 Using this technique, Mostaert et. al. characterized the nanoscale mechanical response of amyloid fibrils discovered in the attachment adhesive of the algae P. linearis.14 By fitting the entropic elastic response of adhesion peaks with the WLC model, they determined the mean persistence length from 100 curves to be 0.34 ± 0.18 nm, which is consistent with proteins such as titin (0.4 nm) and tenascin (0.42 nm).15 The same authors obtained similar results for K. flaccidum and proposed the amyloid quaternary structures as a possible generic mechanism for mechanical strength in natural adhesives.15

Force spectroscopy data of mature amyloid fibrils of the peptide glucagon was analyzed with a WLC model fit to obtain an average persistence length of 0.70 ± 0.15 nm, about twice as rigid of that found in natural adhesives.90 Importantly, even after 1000 cycles, the glucagon extension remained fully reversible, and the fibrils could withstand forces up to 250 pN with the absence of unzipping behavior.90 This is in stark contrast to similar studies of β-lactoglobulin amyloid fibrils which unzip around 30 pN91 and Alzheimer's amyloid β1–40 and β1–42 fibrils that experience fundamental unzipping forces at 33 pN92 and 23 pN,93 respectively. The high tensile strength of the glucagon fibrils can be attributed to the parallel coupling of hydrogen bonds in the peptide's secondary structure amongst which the applied force is evenly distributed.94 Conversely, a study of α-synuclein revealed that fibrils that develop a twisted morphology lack this property and are weaker within the twisted section.95

Statistical Analysis of Shape Fluctuations

Nanomechanics can also be determined from the statistical analysis of the spontaneous thermal fluctuations of fibril geometry since the elastic energy from small strain deformation is on the same order of magnitude as the thermal energy.38 Thermal fluctuations can be monitored by fluorescence imaging96 or tapping-mode AFM [Fig. 2(d)] to determine mechanical97, 98 or morphological98, 99 properties. For instance, statistical analysis of shape fluctuations revealed the persistence length of amyloid fibrils formed by insulin to be 42 ± 30 μm, as compared with 22 ± 3 μm obtained by force spectroscopy.94 Moreover, Figure 3 shows that the analysis of AFM images of the β-lactoglobulin fibrils with the bond correlation function for semiflexible polymers yielded a linear increase of persistence length (1–3 μm) with fibril height (2–6 nm), as expected from the increasing area moment of inertia.43

image
Statistical analysis of tapping mode AFM images to determine mechanical characteristics of various amyloid fibrils. (a) AFM height images of β-lactoglobulin at different magnifications.43 (b) Contour length distribution with a maximum of 5.5 mm and height distribution with peaks corresponding to heights of 2, 4, and 6 nm.43 (c) AFM images of α-lactalbumin, insulin B-chain, β-lactoglobulin, insulin, and TTR(105-115).42 (d) AFM height distributions of corresponding fibrils.42 (e) Analysis of fibril shape with initial tangents horizontally aligned. The calculated bending rigidities for the various types of fibrils are extremely variable, ranging from 1.4 × 10 −28 N m2 for α-lactalbumin to 1.3 × 10 −24 N m2 for TTR(105-115), suggesting a strong correlation between bending rigidity and area moment of inertia.42 Reproduced from a, b, Ref. 43, with permission from NPG; c–e, Ref. 42, with permission from AAAS.

Conversely, the analysis of the geometric fluctuations of amyloid-like protofibrils formed from the N-terminal domain of the E. coli protein HypF demonstrated that fibrils from the same sample with similar heights and widths formed two distinct populations: the first with a persistence length of 69 ± 3 nm and Young's modulus of 500 MPa and the second with a persistence length of 10 ± 1 nm and modulus of 60 MPa.100 This study suggests that while protofibrils are not as strong as mature amyloid fibrils, the prefibrillar aggregates display the same structural heterogeneity as their fibrillar counterparts. Mature amyloid fibrils of α-synuclein were also found to be present in two populations: fibrils that displayed a periodicity of 46 nm and a Young's modulus of 16 GPa and nonperiodic fibrils with a modulus of 24.7 GPa.101 Moreover, ovalbumin was observed to simultaneously assemble into three distinct types of amyloid fibrils of low, intermediate, and high flexibility characterized by persistence lengths of 3 μm, 300 nm, and 63 nm, respectively.85 Additionally, by modeling disorder of bovine insulin amyloid fibrils through thermal fluctuations and structural defects, it could be determined that there is an energy gain of 310 kBT per μm of fibril length upon the lateral interaction of two protofilaments into a mature fibril,102 an interfiber attraction energy much greater than that found in HbS fibers (4 ± 3 kB T/μm),103 indicative of the strong interactions that result in the characteristic mechanical stability of amyloid fibrils.

Other Methods

The nanomechanics of amyloid fibrils have also been determined by other methods, such as electron cryo-microscopy104 and X-ray diffraction.105, 106 Furthermore, several in silico studies have been conducted to model the response of amyloid fibrils subjected to tensile loading,107-109 compressive loading,110, 111 and lateral loading.112 These studies show strong agreement to experimental data and suggest the large values of Young's modulus for amyloid fibrils result from the highly ordered network of densely packed hydrogen bonds found in the fibril core. Figure 4 shows a comparison of various methods used to measure the moduli of the amyloidogenic proteins insulin and α-synuclein.

image
Young's moduli of single insulin and α-synuclein amyloid fibrils according to various AFM-based methods of measurement.42, 79, 83, 84, 94, 101

DISCUSSION OF AMYLOID MATERIAL PROPERTIES

Figure 5 shows that the stiffness of amyloid fibrils as measured by the Young's modulus (E = 100 MPa – 10 GPa) is comparable to the strongest proteinaceous materials known, including microtubules (E = 20 MPa–1.2 GPa),38, 113, 114 collagen (E = 1–9 GPa),115-117 keratin (E = 1.4–8 GPa),115, 118-120 and silk (E = 1 – 10 GPa).121, 122 The structural similarities of dragline silk and amyloid fibrils have been recognized by several groups,123, 124 who have concluded that their shared strength is a result of their characteristic secondary structure consisting of stacked β-sheets that are stabilized by densely packed hydrogen bonds.

image
Comparison of the stiffness of various classes of biological materials according to their corresponding Young's moduli. Amyloid fibrils are among the stiffest proteinaceous materials known and their elastic moduli are comparable to those of keratin, collagen, silk.56, 66, 113-115, 130

A key application of nanomechanical measurements of amyloid fibrils is the elucidation of the role of intermolecular interactions in their core, which give rise to their robust mechanical properties. As the Young's modulus is measured in units of pressure, or equivalently, energy per unit volume, it can be considered as an energy density of interactions within the fibrils. Therefore, the nanomechanical properties of amyloid fibrils are closely connected to the nature of intermolecular interactions, the strongest of which are hydrogen bonds.125, 126 It has been shown that the elastic moduli of amyloid fibrils approach the theoretical limit (on the order of 10 – 20 GPa) of material performance according to the highest density of hydrogen bonds that can be achieved in proteins.42, 127-129 In agreement with this idea, a coarse-grained model representing a single fibril as a string of rigid monomers linked by a simple Gaussian network of hydrogen bonds accurately predicted the experimental elastic moduli, indicating that the hydrogen-bonding network dominates main chain interactions.42 This work supports the conclusion that the densely packed hydrogen-bonding network of the characteristic cross-β core structure is an important consideration in determining the mechanical strength of amyloid materials.

While intermolecular backbone interactions are fundamental to the characteristic structure of amyloid materials, the sequence of side chains that comprise the peptide's primary structure can also have a major effect on the stability of the cross-β core structure and the self-assembly of fibrils. For instance, π-stacking interactions from aromatic residues are thought to contribute favorably energetically and provide directionality and order during the process of self-assembly.131, 132 During aggregation, strong connections form between fibrils from the extensive hydrophobic and van der Waals interactions that occur between functional groups on the outer surface of the fibrils.133 Thus, the highly ordered hierarchical self-assembly of macroscopic fibrillar materials enables the conservation of the desirable nanomechanical properties found in individual fibrils.

Although many amyloids are pathogenic, the material properties of amyloid fibrils have been exploited in several natural functional uses, such as adhesives in the algae P. linearis and K. flaccidum,14, 15 bacterial coatings in E. coli,18 and many natural biofilms.17, 134 The protein TasA, which is the primary structural component of the extracellular matrix in B. subtilis biofilms, has been shown to form amyloid fibrils that are thought to provide structural integrity and bind the cells together.135 Amyloids have also been found in functional uses in humans. For instance, amyloid fibrils of the protein Pmel17 act as a template for melanin polymerization within melanosomes,136, 137 and amyloid-like fibrils have been shown to stabilize and store peptide hormones in pituitary secretory granules.19 In these applications, the amyloids are isolated (in melanosomes or granules), and processes involving their aggregation are likely highly regulated to prevent a cytotoxic response.

AMYLOID FIBRILS AS THE BASIS FOR ARTIFICIAL MATERIALS

The natural functional uses of amyloid fibrils in conjunction with their nanomechanical properties have led researchers to explore their use in artificial materials. Additionally, many groups have used biomimetic techniques to develop de novo amyloid-like peptide platforms with many beneficial applications.

Nonbiological Applications

Recent research on conducting materials for nanoelectronic devices has focused on molecular self-assembly as a bottom-up approach in lieu of traditional top-down fabrication techniques. Amyloid fibrils have shown potential in this field due to their intrinsic propensity to self-assemble into nanostructures with desirable mechanical properties. For example, amyloid fibrils have been used as sacrificial biotemplates to construct metal nanowires from diphenylalanine [Fig. 6(a)],138 insulin,139 Sup35p,140 and hen egg white lysozyme.141 Nanowires that incorporate amyloid peptides as part of the final structure have also been generated by coating the fibril with a conductive conjugated polymer142 or forming trilayered (metal-peptide-metal) nanowires.143 Amyloid fibrils have similarly been used in the fabrication of photovoltaic devices by acting as a template for donor-acceptor materials [Fig. 6(b)].144-146

image
Applications of amyloid and amyloid-like peptides in artificial materials. (a) Schematic representation of the formation of silver nanowires by the reduction of silver ions and subsequent enzymatic degradation of nanofibers (top). Transmission electron microscope images of silver nanowires before (bottom left) and after (bottom center and right) the degradation of the peptide scaffold.138 (b) AFM image of amyloid fibrils decorated with titanium dioxide to form nanowires for applications in hybrid photovoltaic devices.144 (c) Schematic representation of the cysteine conjugation and attachment of fibrils to unmodified gold surfaces.148 (d) SEM images of graphene-amyloid hybrid nanocomposites in a 1:2 ratio; inset depicts macroscopic films with enzyme sensing properties. 151 (e) Self-assembling peptide nanofiber scaffolds from RADA 16-I for brain repair and axon regeneration. 175 (f) Scanning electron micrograph of a cluster of three adult mouse neural stem cells (white circle) embedded in a self-assembling RADA16-BMHP1 designer peptide scaffold. 180 Reproduced from: a, Ref. 139, with permission from AAAS; b, Ref. 145, with permission from Wiley; c, Ref. 149, with permission from ACS; d, Ref. 152, with permission from NPG; e, Ref. 176, with permission from NAS; f, Ref. 181, with permission from Gelain et. al.

Since significant alterations to the primary structure of amyloid peptides to add functionalities are likely to hamper their ability to self-assemble, a recent study reported the synthesis of structure-controllable amyloid peptides (SCAPs) terminated with a three amino acid sequence (either triple lysine or triple glutamic acid) from the amyloidogenic fragment of transthyretin (TTR).147 By mixing various SCAPs, including those that self-assemble and those that are modified with probes and functionalities, nanowires with the desired properties could be formed.

Biological Applications

In addition to these nonbiological applications, amyloid and amyloid-like fibrils have been used in many biological applications. Amyloid fibrils functionalized with small sulfur-containing molecules are able to covalently and irreversibly attach to gold surfaces for applications in biosensing [Fig. 6(c)].148 Other studies have reported the potential use of diphenylalanine nanotubes attached to electrodes149, 150 and hybrid amyloid nanocomposites [Fig. 6(d)]151, 152 for enzyme-biosensor applications. Amyloids have also shown potential in the field of drug delivery by allowing the sustained, controlled release of biologically active peptides through the enzymatic degradation of the amyloid termini.153, 154 The well-defined nanofibrous morphology of hydrogels,155-157 scaffolds,158-160 and films133 derived from amyloid fibrils is structurally similar to that of the extracellular matrix, making amyloid bionanomaterials advantageous for applications in tissue engineering and regenerative medicine.

The self-assembly of amyloid peptides into hierarchical nanostructures in conjunction with their ease of functionalization has given rise to the de novo design of nontoxic amyloid inspired protein nanofibrils.161-165 Under physiological conditions, designer peptide scaffolds undergo spontaneous self-assembly into nanofibrous hydrogel structures rich in β-sheets that have the potential to be used in various delivery systems. For example, the sustained delivery of growth factors to the myocardium using self-assembling peptide scaffolds has been shown to decrease cardiomyocyte morbidity and improve systolic function after myocardial infarction.166, 167 The release profiles and kinetics from self-assembling peptide hydrogels have also been studied for a variety of functional proteins,168 hydrophobic anticancer drugs,169 and various dyes molecularly similar to drug compounds,170 indicative of the versatility of amyloid-like fibrils in controlled delivery systems.

Since designer self-assembling peptide scaffolds are amenable to modifications, they can be easily tailored to a variety of applications. These scaffolds have been shown to promote angiogenesis,171, 172 accelerate wound healing,173 stimulate brain reconstruction174 and axon regeneration [Fig. 6(e)],175 and even enhance dental enamel remineralization.176 The results of these studies, among others, have demonstrated the value of self-assembling scaffolds in the field of regenerative medicine. Designer peptides also have the ability to self-assemble into scaffolds in situ to create nanofibrous microenvironments, enabling an injectable method of tissue regeneration.177 Moreover, self-assembling peptide systems have been beneficial in the fabrication of 3-D cell culture scaffolds by stimulating the attachment and survival of osteoblasts178, 179 and neural cells [Fig. 6(f)].180, 181 The promising results of these studies suggest that self-assembling designer peptide platforms will continue to positively impact the field biotechnology as a result of their favorable morphological and nanomechanical properties.

CONCLUSIONS

In this article, we have discussed the nanoscale mechanics of amyloid fibrils and the possibilities offered by polymer physics inspired concepts and approaches in this area. We have highlighted experimental strategies that allow insights to be gained into the material properties of amyloid fibrils and have discussed how these relate to other materials of biological origin. Finally, we have focused on the key role that materials science can play in guiding the generation of new functionalities and artificial structures from protein building blocks.

ACKNOWLEDGMENTS

The authors acknowledge support from the Whitaker International Program, the Institute of International Education, the Frances and Augustus Newman Foundation, and the Biotechnology and Biological Sciences Research Council.

    Biographies

    • image

      Lisa Volpatti received her Bachelor of Science in Chemical Engineering from the University of Pittsburgh in May 2013. She was awarded a National Science Foundation Graduate Research Fellowship and Whitaker International Fellowship and is currently pursuing a Master of Philosophy in Chemistry at the University of Cambridge. Her interests focus on artificial protein materials and their applications in biotechnology.

    • image

      Tuomas Knowles is a University Reader in Physical Chemistry at the University of Cambridge. He obtained his Master's degree in Physics from the Swiss Federal Institute of Technology in 2004 and his PhD in biophysics from the Cavendish Laboratory in 2008. As a St John's College Junior Research Fellow, he spent time in Cambridge and Harvard, and joined the faculty of the Chemistry Department in Cambridge in 2010. He is the recipient of a number of prizes including the 2012 Royal Society of Chemistry Harrison Meldola Award and an ERC starting grant award.

      The full text of this article hosted at iucr.org is unavailable due to technical difficulties.