Characterization of the protein unfolding processes induced by urea and temperature

Correct folding is critical for the biological activities of proteins. As a contribution to a better understanding of the protein (un)folding problem, we studied the effect of temperature and of urea on peptostreptococcal Protein L destructuration. We performed standard molecular dynamics simulations at 300 K, 350 K, 400 K, and 480 K, both in 10 M urea and in water. Protein L followed at least two alternative unfolding pathways. Urea caused the loss of secondary structure acting preferentially on the β-sheets, while leaving the α-helices almost intact; on the contrary, high temperature preserved the β-sheets and led to a complete loss of the α-helices. These data suggest that urea and high temperature act through different unfolding mechanisms, and protein secondary motives reveal a differential sensitivity to various denaturant treatments. As further validation of our results, replica-exchange molecular dynamics simulations of the temperature-induced unfolding process in the presence of urea were performed. This set of simulations allowed us to compute the thermodynamical parameters of the process and confirmed that, in the configurational space of Protein L unfolding, both of the above pathways are accessible, although to a different relative extent.     

Circular dichroism and the secondary structure of the ROF2 protein from Arabidopsis thaliana

The protein ROF2 from the plant Arabidopsis thaliana acts as a heat stress modulator, being involved in the long-term acquired thermotolerance of the plant. Here we investigate the relationship between the biological function and the structure of ROF2, inferred by circular dichroism (CD) spectroscopy. The far-UV CD spectra, analyzed with the CDPro and DICHROWEB program packages, yield the percentages of α-helices, β-sheets, unordered regions, turns and poly(Pro)II-helices in the secondary structure of ROF2. According to the analysis, the percentages of the structural elements of ROF2 are about 40% for β-sheets, 30% for unordered regions, 17% for turns, 10% for poly(Pro)II-helices and 3% for α-helices. The near-UV CD spectra suggest that ROF2 proteins can associate, forming super-secondary structures. Our CD experiments performed at temperatures between 5 °C and 97 °C indicate that the thermal denaturation of ROF2 caused by a raise in temperature up to 55 °C is followed by a thermal refolding of the protein as the temperature is raised further. The new secondary structure, acquired around 65 °C, remains stable up to 97 °C. The structural stability of ROF2 at high temperatures might play an important role in the experimentally observed thermotolerance of Arabidopsis thaliana.     

Misfolded Amyloid Ion Channels Present Mobile β-Sheet Subunits in Contrast to Conventional Ion Channels

In Alzheimer's disease, calcium permeability through cellular membranes appears to underlie neuronal cell death. It is increasingly accepted that calcium permeability involves toxic ion channels. We modeled Alzheimer's disease ion channels of different sizes (12-mer to 36-mer) in the lipid bilayer using molecular dynamics simulations. Our Aβ channels consist of the solid-state NMR-based U-shaped β-strand-turn-β-strand motif. In the simulations we obtain ion-permeable channels whose subunit morphologies and shapes are consistent with electron microscopy/atomic force microscopy. In agreement with imaged channels, the simulations indicate that β-sheet channels break into loosely associated mobile β-sheet subunits. The preferred channel sizes (16- to 24-mer) are compatible with electron microscopy/atomic force microscopy-derived dimensions. Mobile subunits were also observed for β-sheet channels formed by cytolytic PG-1 β-hairpins. The emerging picture from our large-scale simulations is that toxic ion channels formed by β-sheets spontaneously break into loosely interacting dynamic units that associate and dissociate leading to toxic ionic flux. This sharply contrasts intact conventional gated ion channels that consist of tightly interacting α-helices that robustly prevent ion leakage, rather than hydrogen-bonded β-strands. The simulations suggest why conventional gated channels evolved to consist of interacting α-helices rather than hydrogen-bonded β-strands that tend to break in fluidic bilayers. Nature designs folded channels but not misfolded toxic channels.     

Analysis and prediction of the packing of α-helices against a β-sheet in the tertiary structure of globular proteins

The packing of α-helices and β-sheets in six $\text{α}\text{β}$ proteins (e.g. flavodoxin) has been analysed. The results provide the basis for a computer algorithm to predict the tertiary structure of an $\text{α}\text{β}$ protein from its amino acid sequence and actual assignment of secondary structure.The packing of an individual α-helix against a β-sheet generally involves two adjacent ± 4 rows of non-polar residues on the α-helix at the positions i, i + 4, i + 8, i + 1, i + 5, i + 9. The pattern of interacting β-sheet residues results from the twisted nature of the sheet surface and the attendant rotation of the side-chains. At a more detailed level, four of the α-helical residues (i + 1, i + 4, i + 5 and i + 8) form a diamond that surrounds one particular β-sheet residue, generally isoleucine, leucine or valine. In general, the α-helix sits 10 Å above the sheet and lies parallel to the strand direction.The prediction follows a combinational approach. First, a list of possible β-sheet structures (10⁶ to 10¹⁴) is constructed by the generation of all β-sheet topologies and β-strand alignments. This list is reduced by constraints on topology and the location of non-polar residues to mediate the sheet/helix packing, and then rank-ordered on the extent of hydrogen bonding. This algorithm was uniformly applied to 16 $\text{α}\text{β}$ domains in 13 proteins. For every structure, one member of the reduced list was close to the crystal structure; the root-mean-square deviation between equivalenced C^α atoms averaged 5.6 Å for 100 residues. For the $\text{α}\text{β}$ proteins with pure parallel β-sheets, the total number of structures comparable to or better than the native in terms of hydrogen bonds was between 1 and 148. For proteins with mixed β-sheets, the worst case is glyceraldehyde-3-phosphate dehydrogenase, where as many as 3800 structures would have to be sampled. The evolutionary significance of these results as well as the potential use of a combinatorial approach to the protein folding problem are discussed.     

BCL::Score—Knowledge Based Energy Potentials for Ranking Protein Models Represented by Idealized Secondary Structure Elements

The topology of most experimentally determined protein domains is defined by the relative arrangement of secondary structure elements, i.e. α-helices and β-strands, which make up 50–70% of the sequence. Pairing of β-strands defines the topology of β-sheets. The packing of side chains between α-helices and β-sheets defines the majority of the protein core. Often, limited experimental datasets restrain the position of secondary structure elements while lacking detail with respect to loop or side chain conformation. At the same time the regular structure and reduced flexibility of secondary structure elements make these interactions more predictable when compared to flexible loops and side chains. To determine the topology of the protein in such settings, we introduce a tailored knowledge-based energy function that evaluates arrangement of secondary structure elements only. Based on the amino acid C_β atom coordinates within secondary structure elements, potentials for amino acid pair distance, amino acid environment, secondary structure element packing, β-strand pairing, loop length, radius of gyration, contact order and secondary structure prediction agreement are defined. Separate penalty functions exclude conformations with clashes between amino acids or secondary structure elements and loops that cannot be closed. Each individual term discriminates for native-like protein structures. The composite potential significantly enriches for native-like models in three different databases of 10,000–12,000 protein models in 80–94% of the cases. The corresponding application, “BCL::ScoreProtein,” is available at www.meilerlab.org.     

On the presence of short-range periodicities in protein structures that are not related to established secondary structure elements

Standard secondary structure elements such as α-helices or β-sheets, are characterized by repeating backbone torsion angles (φ,ψ) at the single residue level. Two-residue motifs of the type (φ,ψ)₂ are also observed in nonlinear conformations, mainly turns. Taking these observations a step further, it can be argued that there is no a priori reason why the presence of higher order periodicities can not be envisioned in protein structures, such as, for example, periodic transitions between successive residues of the type (…-α-β-α-β-α-…), or (…-β-α_L-β-α_L-β-…), or (…-α-β-α_L-α-β-α_L-…), and so forth, where the symbols (α,β,α_L) refer to the established Ramachandran-based residue conformations. From all such possible higher order periodicities, here we examine the deposited (with the PDB) protein structures for the presence of short-range periodical conformations comprising five consecutive residues alternating between two (and only two) distinct Ramachandran regions, for example, conformations of the type (α-β-α-β-α) or (β-α_L-β-α_L-β), and so forth. Using a probabilistic approach, we have located several thousands of such peptapeptides, and these were clustered and analyzed in terms of their structural characteristics, their sequences, and their putative functional correlations using a gene ontology-based approach. We show that such nonstandard short-range periodicities are present in a large and functionally diverse sample of proteins, and can be grouped into two structurally conserved major types. Examination of the structural context in which these peptapeptides are observed gave no conclusive evidence for the presence of a persistent structural or functional role of these higher order periodic conformations.     

High temperature unfolding of Bacillus anthracis amidase-03 by molecular dynamics simulations

The stability of amidase-03 structure (a cell wall hydrolase protein) from Bacillus anthracis was studied using classical molecular dynamics (MD) simulation. This protein (GenBank accession number: {"type":"entrez-protein","attrs":{"text":"NP_844822","term_id":"30262445","term_text":"NP_844822"}}NP_844822) contains an amidase-03 domain which is known to exhibit the catalytic activity of N-acetylmuramoyl-L-alanine amidase (digesting MurNAc-Lalanine linkage of bacterial cell wall). The amidase-03 enzyme has stability at high temperature due to the core formed by the combination of several secondary structure elements made of β-sheets. We used root-mean-square-displacement (RMSD) of the simulated structure from its initial state to demonstrate the unfolding of the enzyme using its secondary structural elements. Results show that amidase-03 unfolds in transition state ensemble (TSE). The data suggests that α-helices unfold before β-sheets from the core during simulation.     

Fast in silico protein folding by introduction of alternating hydrogen bond potentials

We accelerate protein folding in all-atom molecular dynamics simulations by introducing alternating hydrogen bond potentials as a supplement to the force field. The alternating hydrogen bond potentials result in accelerated hydrogen bond reordering, which leads to rapid formation of secondary structure elements. The method does not require knowledge of the native state but generates the potentials based on the development of the tertiary structure in the simulation. In protein folding, the formation of secondary structure elements, especially α-helix and β-sheet, is very important, and we show that our method can fold both efficiently and with great speed.     

Folding Simulations of a De Novo Designed Protein with a βαβ Fold

βαβ structural motifs are commonly used building blocks in protein structures containing parallel β-sheets. However, to our knowledge, no stand-alone βαβ structure has been observed in nature to date. Recently, for the first time that we know of, a small protein with an independent βαβ structure (DS119) was successfully designed in our laboratory. To understand the folding mechanism of DS119, in the study described here, we carried out all-atom molecular dynamics and coarse-grained simulations to investigate its folding pathways and energy landscape. From all-atom simulations, we successfully observed the folding event and got a stable folded structure with a minimal root mean-square deviation of 2.6 Å with respect to the NMR structure. The folding process can be described as a fast collapse phase followed by rapid formation of the central helix, and then slow formation of a parallel β-sheet. By using a native-centric Gō-like model, the cooperativity of the system was characterized in terms of the calorimetric criterion, sigmoidal transitions, conformation distribution shifts, and free-energy profiles. DS119 was found to be an incipient downhill folder that folds more cooperatively than a downhill folder, but less cooperatively than a two-state folder. This may reflect the balance between the two structural elements of DS119: the rapidly formed α-helix and the slowly formed parallel β-sheet. Folding times estimated from both the all-atom simulations and the coarse-grained model were at microsecond level, making DS119 another fast folder. Compared to fast folders reported previously, DS119 is, to the best of our knowledge, the first that exhibits a parallel β-sheet.     

Induced Secondary Structure and Polymorphism in an Intrinsically Disordered Structural Linker of the CNS: Solid-State NMR and FTIR Spectroscopy of Myelin Basic Protein Bound to Actin

The 18.5 kDa isoform of myelin basic protein (MBP) is a peripheral membrane protein that maintains the structural integrity of the myelin sheath of the central nervous system by conjoining the cytoplasmic leaflets of oligodendrocytes and by linking the myelin membrane to the underlying cytoskeleton whose assembly it strongly promotes. It is a multifunctional, intrinsically disordered protein that behaves primarily as a structural stabilizer, but with elements of a transient or induced secondary structure that represent binding sites for calmodulin or SH3-domain-containing proteins, inter alia. In this study we used solid-state NMR (SSNMR) and Fourier transform infrared (FTIR) spectroscopy to study the conformation of 18.5 kDa MBP in association with actin microfilaments and bundles. FTIR spectroscopy of fully ¹³C,¹⁵N-labeled MBP complexed with unlabeled F-actin showed induced folding of both protein partners, viz., some increase in β-sheet content in actin, and increases in both α-helix and β-sheet content in MBP, albeit with considerable extended structure remaining. Solid-state NMR spectroscopy revealed that MBP in MBP-actin assemblies is structurally heterogeneous but gains ordered secondary structure elements (both α-helical and β-sheet), particularly in the terminal fragments and in a central immunodominant epitope. The overall conformational polymorphism of MBP is consistent with its in vivo roles as both a linker (membranes and cytoskeleton) and a putative signaling hub.     

Union of Geometric Constraint-Based Simulations with Molecular Dynamics for Protein Structure Prediction

Although proteins are a fundamental unit in biology, the mechanism by which proteins fold into their native state is not well understood. In this work, we explore the assembly of secondary structure units via geometric constraint-based simulations and the effect of refinement of assembled structures using reservoir replica exchange molecular dynamics. Our approach uses two crucial features of these methods: i), geometric simulations speed up the search for nativelike topologies as there are no energy barriers to overcome; and ii), molecular dynamics identifies the low free energy structures and further refines these structures toward the actual native conformation. We use eight α-, β-, and α/β-proteins to test our method. The geometric simulations of our test set result in an average RMSD from native of 3.7 Å and this further reduces to 2.7 Å after refinement. We also explore the question of robustness of assembly for inaccurate (shifted and shortened) secondary structure. We find that the RMSD from native is highly dependent on the accuracy of secondary structure input, and even slightly shifting the location of secondary structure along the amino acid sequence can lead to a rapid decrease in RMSD to native due to incorrect packing.     

Physicochemical and immunological characterization of the type E botulinum neurotoxin binding protein purified fromClostridium botulinum

Type E botulinum neurotoxin is produced byClostridium botulinum along with a neurotoxin binding protein which helps protect the neurotoxin from adversepH, temperature, and proteolytic conditions. The neurotoxin binding protein has been purified as a 118-kDa protein. Secondary structure content of the neurotoxin binding protein as revealed by far-UV circular dichroism spectroscopy was 19% α-helix, 50%β-sheets, 28% random coils, and 3%β-turns. This compared to 22% α-helix, 44%β-sheets, 34% random coils, and noβ-turns of the type E botulinum neurotoxin. The complex of the two proteins revealed 25%α-helix, 45%β-sheets, 27% random coils, and 3%β-turns, suggesting a significant alteration at least in theα-helical folding of the two proteins upon their interaction. Tyrosine topography is altered considerably (28%) when the neurotoxin and its binding protein are separated, indicating strong interaction between the two proteins. Gel filtration results suggested that type E neurotoxin binding protein clearly complexes with type E neurotoxin. The interaction is favored at lowpH as indicated by an initial binding rate of 8.4 min^?1 atpH 5.7 compared to 4.0 min^?1 atpH 7.5 as determined using a fiber optic-based biosensor. The neurotoxin and its binding protein apparently are of equivalent antigenicity, as both reacted equally on enzyme-linked immunosorbent assay to polyclonal antibodies raised against the toxoid of their complex.     

Estimating loop length from CryoEM images at medium resolutions

Background

De novo protein modeling approaches utilize 3-dimensional (3D) images derived from electron cryomicroscopy (CryoEM) experiments. The skeleton connecting two secondary structures such as α-helices represent the loop in the 3D image. The accuracy of the skeleton and of the detected secondary structures are critical in De novo modeling. It is important to measure the length along the skeleton accurately since the length can be used as a constraint in modeling the protein.

Results

We have developed a novel computational geometric approach to derive a simplified curve in order to estimate the loop length along the skeleton. The method was tested using fifty simulated density images of helix-loop-helix segments of atomic structures and eighteen experimentally derived density data from Electron Microscopy Data Bank (EMDB). The test using simulated density maps shows that it is possible to estimate within 0.5Å of the expected length for 48 of the 50 cases. The experiments, involving eighteen experimentally derived CryoEM images, show that twelve cases have error within 2Å.

Conclusions

The tests using both simulated and experimentally derived images show that it is possible for our proposed method to estimate the loop length along the skeleton if the secondary structure elements, such as α-helices, can be detected accurately, and there is a continuous skeleton linking the α-helices.

     

BCL::Fold - De Novo Prediction of Complex and Large Protein Topologies by Assembly of Secondary Structure Elements

Computational de novo protein structure prediction is limited to small proteins of simple topology. The present work explores an approach to extend beyond the current limitations through assembling protein topologies from idealized α-helices and β-strands. The algorithm performs a Monte Carlo Metropolis simulated annealing folding simulation. It optimizes a knowledge-based potential that analyzes radius of gyration, β-strand pairing, secondary structure element (SSE) packing, amino acid pair distance, amino acid environment, contact order, secondary structure prediction agreement and loop closure. Discontinuation of the protein chain favors sampling of non-local contacts and thereby creation of complex protein topologies. The folding simulation is accelerated through exclusion of flexible loop regions further reducing the size of the conformational search space. The algorithm is benchmarked on 66 proteins with lengths between 83 and 293 amino acids. For 61 out of these proteins, the best SSE-only models obtained have an RMSD100 below 8.0 Å and recover more than 20% of the native contacts. The algorithm assembles protein topologies with up to 215 residues and a relative contact order of 0.46. The method is tailored to be used in conjunction with low-resolution or sparse experimental data sets which often provide restraints for regions of defined secondary structure.     

Distributions of amino acids suggest that certain residue types more effectively determine protein secondary structure

Exponential growth in the number of available protein sequences is unmatched by the slower growth in the number of structures. As a result, the development of efficient and fast protein secondary structure prediction methods is essential for the broad comprehension of protein structures. Computational methods that can efficiently determine secondary structure can in turn facilitate protein tertiary structure prediction, since most methods rely initially on secondary structure predictions. Recently, we have developed a fast learning optimized prediction methodology (FLOPRED) for predicting protein secondary structure (Saraswathi et al. in JMM 18:4275, 2012). Data are generated by using knowledge-based potentials combined with structure information from the CATH database. A neural network-based extreme learning machine (ELM) and advanced particle swarm optimization (PSO) are used with this data to obtain better and faster convergence to more accurate secondary structure predicted results. A five-fold cross-validated testing accuracy of 83.8 % and a segment overlap (SOV) score of 78.3 % are obtained in this study. Secondary structure predictions and their accuracy are usually presented for three secondary structure elements: α-helix, β-strand and coil but rarely have the results been analyzed with respect to their constituent amino acids. In this paper, we use the results obtained with FLOPRED to provide detailed behaviors for different amino acid types in the secondary structure prediction. We investigate the influence of the composition, physico-chemical properties and position specific occurrence preferences of amino acids within secondary structure elements. In addition, we identify the correlation between these properties and prediction accuracy. The present detailed results suggest several important ways that secondary structure predictions can be improved in the future that might lead to improved protein design and engineering.     

Induced β-Barrel Formation of the Alzheimer's Aβ25-35 Oligomers on Carbon Nanotube Surfaces: Implication for Amyloid Fibril Inhibition

Recent experimental studies show that carbon nanotubes impact the aggregation process of proteins associated with neurodegenerative diseases. However, the details of molecular interactions between proteins and carbon nanotubes are still not well understood. In this study, we investigate the initial adsorption features and dynamics of the Alzheimer's amyloid-β peptide spanning residues 25-35 (Aβ25-35) on a single-walled carbon nanotube (SWNT) surface using fully atomic molecular dynamics simulations (MD) in explicit solvent. The initial configurations of the Aβ25-35 peptides consist of two preformed bilayer β-sheets, each with four or five β-strands in parallel or mixed antiparallel-parallel orientations. Our simulations show, for what we believe is the first time, that two disjointed Aβ25-35 β-sheets with mixed antiparallel-parallel strands can assemble into β-barrels wrapping the SWNT. In contrast, both simulations of Aβ25-35 without SWNT, and simulations of SWNT−Aβ25-35 with purely parallel β-strands, lead to disordered aggregates. We find that Aβ25-35 β-barrel formation involves at least two steps: i), curving of the Aβ25-35 β-sheets as a result of strong hydrophobic interactions with carbon nanotube concomitantly with dehydration of the SWNT-peptide interface; and ii), intersheet backbone hydrogen bond formation with fluctuating intrasheet hydrogen bonds. Detailed analysis of the conversion shows that β-barrel formation on SWNT surface results from the interplay of dehydration and peptide-SWNT/peptide-peptide interactions. Implications of our results on amyloid fibril inhibition are discussed.     

A step-by-step classification algorithm of protein secondary structures based on double-layer SVM model

In this paper, a step-by-step classification algorithm based on double-layer SVM model is constructed to predict the secondary structure of proteins. The most important feature of this algorithm is to improve the prediction accuracy of α+β and α/β classes through transforming the prediction of two classes of proteins, α+β and α/β classes, with low accuracy in the past, into the prediction of all-α and all-β classes with high accuracy. A widely-used dataset, 25PDB dataset with sequence similarity lower than 40%, is used to evaluate this method. The results show that this method has good performance, and on the basis of ensuring the accuracy of other three structural classes of proteins, the accuracy of α+β class proteins is improved significantly.     

Refinement of protein-protein complexes in contact map space with metadynamics simulations

Accurate protein-protein complex prediction, to atomic detail, is a challenging problem. For flexible docking cases, current state-of-the-art docking methods are limited in their ability to exhaustively search the high dimensionality of the problem space. In this study, to obtain more accurate models, an investigation into the local optimization of initial docked solutions is presented with respect to a reference crystal structure. We show how physics-based refinement of protein-protein complexes in contact map space (CMS), within a metadynamics protocol, can be performed. The method uses 5 times replicated 10 ns simulations for sampling and ranks the generated conformational snapshots with ZRANK to identify an ensemble of n snapshots for final model building. Furthermore, we investigated whether the reconstructed free energy surface (FES), or a combination of both FES and ZRANK, referred to as CS_α, can help to reduce snapshot ranking error.     

Backbone relaxation coupled to the ionization of internal groups in proteins: a self-guided Langevin dynamics study

Pathways of structural relaxation triggered by ionization of internal groups in staphylococcal nuclease (SNase) were studied through multiple self-guided Langevin dynamics (SGLD) simulations. Circular dichroism, steady-state Trp fluorescence, and nuclear magnetic resonance spectroscopy have shown previously that variants of SNase with internal Glu, Asp, and Lys at positions 66 or 92, and Arg at position 66, exhibit local reorganization or global unfolding when the internal ionizable group is charged. Except for Arg-66, these internal ionizable groups have unusual pK_a values and are neutral at physiological pH. The structural trends observed in the simulations are in general agreement with experimental observations. The I92D variant, which unfolds globally upon ionization of Asp-92, in simulations often exhibits extensive hydration of the protein core, and sometimes also significant perturbations of the β-barrel. In the crystal structure of the V66R variant, the β1 strand from the β-barrel is domain-swapped; in the simulations, the β1 strand is sometimes partially released. The V66K variant, which in solutions shows reorganization of six residues at the C-terminus of helix α1 and perturbations in the β-barrel structure, exhibits fraying of three residues of helix α1 in one simulation, and perturbations and partial unfolding of three β-strands in a few other simulations. In sharp contrast, very small structural changes were observed in simulations of the wild-type protein. The simulations indicate that charging of internal groups frequently triggers penetration of water into the protein interior. The pK_a values of Asp-92 and Arg-66 calculated with continuum methods on SGLD-relaxed structures reached the normal values in most simulations. Detailed analysis of accuracy and performance of SGLD demonstrates that SGLD outperforms LD in sampling of alternative protein conformations without loss of the accuracy and level of detail characteristic of regular LD.     

Stable Polyglutamine Dimers Can Contain β-Hairpins with Interdigitated Side Chains—But Not α-Helices, β-Nanotubes, β-Pseudohelices,or Steric Zippers

A common thread connecting nine fatal neurodegenerative protein aggregation diseases is an abnormally expanded polyglutamine tract found in the respective proteins. Although the structure of this tract in the large mature aggregates is increasingly well described, its structure in the small early aggregates remains largely unknown. As experimental evidence suggests that the most toxic species along the aggregation pathway are the small early ones, developing strategies to alleviate disease pathology calls for understanding the structure of polyglutamine peptides in the early stages of aggregation. Here, we present a criterion, grounded in available experimental data, that allows for using kinetic stability of dimers to assess whether a given polyglutamine conformer can be on the aggregation path. We then demonstrate that this criterion can be assessed using present-day molecular dynamics simulations. We find that although the α-helical conformer of polyglutamine is very stable, dimers of α-helices lack the kinetic stability necessary to support further oligomerization. Dimers of steric zipper, β-nanotube, and β-pseudohelix conformers are also too short-lived to initiate aggregation. The β-hairpin-containing conformers, instead, invariably form very stable dimers when their side chains are interdigitated. Combining these findings with the implications of recent solid-state NMR data on mature fibrils, we propose a possible pathway for the initial stages of polyglutamine aggregation, in which β-hairpin-containing conformers act as templates for fibril formation.