Polymorphic structures of rapidly twisting 40-residue amyloid-β fibrils

Using cryogenic electron microscopy, this study reveals that rapidly twisting 40-residue amyloid-β fibrils exhibit three distinct polymorphic structures with similar cross-over distances but differing in twist handedness, symmetry, and molecular conformations, some of which resemble previously characterized brain-derived fibrils despite having shorter ordered segments.

The Gist

The Big Picture: The "Twisting Rope" Mystery

Imagine you have a long, twisted rope made of tiny Lego bricks. In the world of biology, these ropes are called amyloid fibrils, and they are formed by a protein called Amyloid-beta (Aβ). When these ropes form in the brain, they are the main culprit behind Alzheimer's disease.

For a long time, scientists thought that if you grew these ropes in a lab, they would all look the same. But this paper reveals a surprising secret: Even when you grow them under the exact same conditions, these ropes can twist into three completely different shapes.

Think of it like baking cookies. If you use the exact same recipe, oven temperature, and baking time, you might expect identical cookies. But in this case, the "dough" (the protein) decided to twist itself into three distinct patterns, just like a rope can be twisted left-handed, right-handed, or in a weird, asymmetrical knot.

The Discovery: The "Fast Twisters"

The researchers focused on a specific type of rope that twists very quickly. If you look at these ropes under a microscope, they look like a corkscrew with a twist every 25 nanometers (that's incredibly small—about the width of 250 atoms).

Usually, these ropes twist much more slowly (every 36 to 100 nanometers). The team wondered: "Do these fast-twisting ropes have a secret, unique structure inside?"

They used a super-powerful camera called Cryo-EM (which takes 3D pictures of frozen proteins) to look inside. They found three distinct "species" of these fast-twisting ropes, all growing side-by-side in the same test tube.

The Three Characters

The team named these three rope types based on how their internal "bricks" (molecules) are arranged:

1. The Symmetrical Twin (RT-Aβ40(21)):

   • The Analogy: Imagine two people holding hands and spinning around a central pole. They are mirror images of each other.
   • The Twist: This rope twists to the right (like a standard screw).
   • The Surprise: It looks very similar to ropes found in the brains of Alzheimer's patients, except the "tail" of the rope is shorter and messier. It's like finding a familiar face, but they are wearing a different hat.

2. The Left-Handed Mirror (RT-Aβ40(C2)):

   • The Analogy: Again, two people holding hands, but this time they are spinning the opposite way (left-handed).
   • The Twist: This is a big deal because scientists previously thought only ropes found in human brains twisted left-handed. Finding one grown in a simple lab solution proves that the brain isn't the only place that can make this shape.
   • The Structure: It has a very specific "herringbone" pattern where the molecules interlock tightly, but the angle is slightly sharper than its slow-twisting cousins, causing it to spiral faster.

3. The Odd One Out (RT-Aβ40(C1)):

   • The Analogy: Imagine two people holding hands, but one is tall and the other is short, or one is wearing a suit and the other is in a t-shirt. They are not mirror images.
   • The Twist: This is the most unique discovery. For the first time, scientists found a wild-type (normal) amyloid rope where the two strands inside are completely different from each other. It's like a zipper where the left side has different teeth than the right side. This has never been seen before in normal amyloid ropes.

Why Do They Twist So Fast?

You might ask, "Why do these ropes twist so tightly compared to the slow ones?"

The researchers found a clever geometric reason. Imagine a spiral staircase.

• If the steps (the ordered part of the protein) are long, the staircase has to be wide to avoid the steps hitting each other. This makes a slow, gentle twist.
• If the steps are short, the staircase can be narrower and twist tighter and faster without the steps crashing into each other.

In these fast-twisting ropes, the "ordered" part of the protein is shorter. Because the building blocks are shorter, the rope can spin around much more quickly without breaking apart.

Why Does This Matter?

This paper is like solving a puzzle about how Alzheimer's works:

1. Nature is Messy: It shows that even in a controlled lab, nature creates variety. There isn't just "one" Alzheimer's rope; there are many shapes.
2. Lab vs. Brain: Some of these lab-grown ropes look almost exactly like ropes found in human brains. This is good news! It means we can study these ropes in the lab to understand the disease, because the lab versions are "real" enough.
3. New Secrets: The discovery of the "Odd One Out" (the asymmetrical rope) opens a new door. It suggests that these proteins are more flexible and complex than we ever imagined.

The Takeaway

Think of Amyloid-beta proteins as a group of dancers. The paper shows that even if you give them the same music and the same dance floor, they can spontaneously form three different dance formations: a symmetrical right-handed spin, a symmetrical left-handed spin, and a weird, asymmetrical shuffle.

Understanding these different "dance moves" helps scientists figure out which ones are the most dangerous to the brain and how to stop them from forming in the first place.

Technical Summary

1. Problem Statement

Amyloid-β (Aβ) fibrils, particularly those formed by the 40-residue peptide (Aβ40), exhibit polymorphism, meaning they can adopt multiple distinct molecular structures depending on growth conditions. While previous studies have characterized various Aβ40 fibril morphologies, a specific class of rapidly twisting fibrils (defined by a "cross-over distance" or period of width modulation of approximately 25 nm) remained structurally uncharacterized at high resolution.

• Knowledge Gap: Existing molecular models for Aβ40 fibrils generally have cross-over distances ranging from 36 nm to over 100 nm. It was unclear whether the shorter 25 nm cross-over distance correlated with novel molecular conformations (potentially resembling S-shaped folds seen in Aβ42) or if these were simply variations of known structures.
• Scientific Question: Do rapidly twisting Aβ40 fibrils represent unique structural polymorphs, and how do their molecular conformations, symmetries, and intermolecular contacts compare to slowly twisting fibrils and those extracted directly from Alzheimer's disease (AD) brain tissue?

2. Methodology

The study employed a combination of in vitro fibril growth, cryogenic electron microscopy (cryo-EM), and computational modeling.

• Sample Preparation: Synthetic Aβ40 peptides were aggregated in vitro under quiescent conditions (no stirring) at 37°C, 200 μM concentration, in 10 mM sodium phosphate buffer (pH 7.4). This specific condition yielded a high population of rapidly twisting fibrils.
• Imaging:
  • Negative Stain TEM: Used for initial screening to identify conditions producing rapidly twisting fibrils.
  • Cryo-EM: High-resolution data was collected on a ThermoFisher Titan Krios G4 microscope (300 keV) equipped with a Gatan K3 camera.
• Image Processing:
  • Data processing was performed using RELION (versions 4.0 and 5.0).
  • Particles were picked using Topaz after training on manually selected particles.
  • 3D Classification was used to separate distinct polymorphs based on symmetry and density features.
  • Helical Reconstruction: Imposed helical symmetry (and specific point symmetries like quasi-21, C2, or C1) to generate 3D density maps.
• Model Building & Refinement:
  • Initial atomic models were built manually in Coot.
  • Models were refined using Xplor-NIH with restraints derived from the cryo-EM density maps, including non-crystallographic symmetry (NCS) and distance symmetry potentials.
  • Handedness Determination: The chirality (left- vs. right-handed twist) was determined by comparing molecular models against density maps at resolutions better than ~3 Å, specifically looking at the alignment of backbone carbonyl oxygens.
• Molecular Dynamics (MD): 200 ns all-atom MD simulations were performed in explicit solvent (100 mM NaCl) to analyze the stability and dynamics of sidechains (e.g., F19, F20, M35) and ion interactions in the rapidly twisting vs. slowly twisting polymorphs.

3. Key Contributions & Results

The study successfully resolved high-resolution structures (2.50 Å – 2.75 Å) for three distinct rapidly twisting Aβ40 polymorphs and one related variant, revealing unexpected structural diversity.

A. Identification of Three Distinct Polymorphs

Despite having similar cross-over distances (~24–32 nm), the three main polymorphs differ significantly in symmetry, twist handedness, and molecular conformation:

1. RT-Aβ40(21):
   • Symmetry: Quasi-21 symmetry (two subunits related by a 2-fold screw axis).
   • Twist: Right-handed.
   • Ordered Core: Residues 15–36. N-terminus (1–14) is disordered.
   • Contacts: Features a "steric zipper" between residues 25–27 and specific hydrophobic contacts (L17, F19, V24 vs. I32, L34, V36).
2. RT-Aβ40(C2):
   • Symmetry: C2 symmetry.
   • Twist: Left-handed.
   • Ordered Core: Residues 17–39.
   • Unique Feature: Contains additional outer density layers (partially disordered β-sheets) covering exposed hydrophobic residues (A30, I32, M35).
   • Contacts: Distinct inter-subunit contacts (F19-L34, F19-V36, V24-I31) compared to RT-Aβ40(21).
3. RT-Aβ40(C1):
   • Symmetry: C1 symmetry (No symmetry between the two cross-β subunits).
   • Twist: Right-handed.
   • Significance: This is the first report of a wild-type Aβ40 fibril polymorph containing two conformationally inequivalent subunits within the same core.
   • Structure: The "lower" subunit (residues 14–37) and "upper" subunit (residues 15–37) have different backbone conformations (e.g., different β-strand lengths) and sidechain orientations (e.g., M35 interactions differ between subunits).

B. Structural Comparisons and Mechanisms of Twisting

• Relation to Known Structures:
  • RT-Aβ40(21) closely resembles structures found in AD brain tissue (meninges) and brain-seeded fibrils (PDB: 8FF2, 8QN6, 8OT4), but with a shorter ordered core (missing residues 1–14 order) and different C-terminal conformations.
  • RT-Aβ40(C2) resembles slowly twisting, brain-seeded fibrils (PDB: 6W0O, 8FF3) but with a shorter ordered segment (starting at L17 instead of H14) and a larger "herringbone" angle between subunits.
• Mechanism of Rapid Twist: The study concludes that the rapid twist (short cross-over distance) is not caused by a single large conformational change at a specific site. Instead, it results from:
  1. Shorter conformationally ordered segments: Geometric constraints suggest that shorter ordered lengths allow for tighter twisting without breaking intermolecular hydrogen bonds.
  2. Subtle conformational variations: Small changes in backbone torsion angles and sidechain orientations distributed across the structure.
• S-Shaped Conformations: Contrary to the hypothesis that rapidly twisting Aβ40 might adopt S-shaped folds (common in Aβ42), no S-shaped conformations were found in these wild-type Aβ40 fibrils.

C. Dynamics and Solvation

• MD Simulations: Revealed that M35 sidechains are dynamically disordered in both rapid and slow polymorphs.
• Ion Binding: Differences in ion distribution (Na+ and Cl-) were observed. In RT-Aβ40(C2), ions associate differently with E22 and K28 compared to the slowly twisting polymorph, potentially influencing electrostatic interactions that modulate twist.

4. Significance

• Polymorphism Complexity: The study demonstrates that a single set of in vitro growth conditions can yield multiple distinct structural polymorphs with similar macroscopic morphologies (cross-over distances) but fundamentally different molecular architectures.
• Symmetry Breaking: The discovery of the C1 asymmetric polymorph (RT-Aβ40(C1)) challenges the assumption that wild-type Aβ fibrils must possess internal symmetry (21 or C2).
• Brain vs. Lab: The findings bridge the gap between in vitro and in vivo structures. They show that in vitro grown fibrils can replicate the right-handed twist and core structures of brain-derived fibrils, suggesting that the "seed" structure can be propagated faithfully, though subtle differences (like N-terminal disorder) exist.
• Structure-Function Relationship: The work provides a geometric and energetic explanation for how cross-over distance correlates with molecular structure, suggesting that the length of the ordered core is a primary determinant of fibril twist rate.
• Implications for AD: Understanding these polymorphic states is crucial for developing structure-specific diagnostics and therapeutics for Alzheimer's disease, as different strains may have varying toxicities and propagation rates.

5. Data Availability

• PDB Codes: 11EN (RT-Aβ40(21)), 11EO (RT-Aβ40(C2)), 11EP (RT-Aβ40(C1)).
• Scripts and Trajectories: Available via the provided DOI link in the manuscript.