30 Jan 2025

Over the past decade, cryo-electron microscopy has afforded scientists an unprecedented view deep into the core of tau fibrils, revealing shapes that track with different neurodegenerative diseases (Oct 2021 news). While the images are stunning, cryo-EM does not explain how these structures formed. In Science Advances January 22, scientists led by Marc Diamond of the University of Texas Southwestern Medical Center in Dallas present a method that distinguishes fibrils from different tauopathies, and pinpoints critical amino acids involved in shaping them.

The scientists did this by substituting each amino acid across tau’s four-repeat domains with a functionally neutral alanine residue, aka alanine scanning, and then used biosensor cell lines to see how each mutant influenced tau’s incorporation into growing fibrils. For each structural strain of tau fibrils, they identified amino acids without which the strain cannot form its signature core fold. In addition, this technique hinted at other parts of the protein—outside of the cryo-EM-resolvable core—that might cooperate in fibril folding.

“This approach is a welcome addition to the technical armamentarium for defining the molecular basis of proteopathic strains, and it furnishes persuasive new support for the key role of prionic mechanisms in tauopathies,” commented Lary Walker of Emory University in Atlanta. “The findings underscore the importance of amino acids in the amyloid core for defining the strain-like variations of tau,” he added. 

First author Jaime Vaquer-Alicea and colleagues used biosensor cell lines developed in Diamond’s lab to probe the basis of different tau assemblies (Oct 2014 news). They reasoned that only tau monomers with a sequence that can conform to the core fibril fold of a given tau strain will be incorporated into it. They used alanine scanning to mutate the length of tau’s four-repeat domains, expressed them in biosensor cell lines, and measured how each incorporates into different tau fibrils.

First, they tested this method on 18 established biosensor lines that had been created with fibrils from different recombinant proteins or human tauopathy cases. All of these express the tau repeat domain, harboring disease-associated P301L and V337M mutations—aka Tau RD (LM). Each propagated a different strain; for example, the DS13 line was established with tau fibrils from a person with corticobasal degeneration.

Vaquer-Alicea found that alanine substitutions within the second or third repeat domains of tau, particularly those that fell within the VQIVYK sequence that forms the core of most tau fibrils, rebuffed incorporation of those monomers into growing tau fibrils. Different incorporation patterns emerged for the different strains, suggesting the technique could distinguish between different tau folds.

The scientists also spotted bona fide tau fibrils within the biosensor cell lines via cryo-EM tomography (image at right). This, and the effects of alanine substitution being strongest around the known core of fibrils, serves up solid evidence that tau fibrils form within the biosensor cell lines, Diamond told Alzforum. The findings refute claims made by a previous study that fibrils can’t form in these cell lines (May 2020 news).

Essay parameters established, Vaquer-Alicea deployed it for fibrils of tau derived from human brain samples. They generated two biosensor strains, in which they expressed FRET-ready pairs of 4R-tau only, or both 3R- and 4R-tau, to accommodate fibrils from mixed tauopathies. FRET, aka fluorescence resonance energy transfer, can be used to monitor growing fibrils. They then seeded these cell lines with fibrils from eight AD, five CBD, two CTE, and six PSP cases, and allowed two days for new fibrils to form within the cells. With fibrils established, the scientists then transduced those cells with alanine variants and monitored their incorporation (image below).

Tau, Incorporated. A library of tau mutants generated by alanine substitution (red residues, left), was transduced into biosensor cell lines containing pre-existing tau fibril strains. The degree of incorporation of each mutant into the growing fibrils (middle) is measured via FRET (right). [Courtesy of Vaquer-Alicea et al., Science Advances, 2025.]

Strikingly, they found that seeds from the different diseases incorporated markedly distinct alanine variants of the RD sequence. Crucial residues for CBD fibrils spanned the beginning of the second repeat to R4 and beyond, while those needed for AD started toward the end of R2 and extended past R4. Notably, hits from the alanine scan corresponded tightly with the cores of each type of fibril that was resolved via cryo-EM.

Within the R3 and R4 domains, AD and CTE required similar residues for incorporation into fibrils, in agreement with their similar core structures as reported by cryo-EM. Interestingly, for CTE only, residues in the beginning of the R2 domain were essential, despite being excluded from the fibril core. In the cryo-EM resolved structure of CTE fibrils, these residues were exposed to solvent. Diamond hypothesized that these residues might somehow play a role in the formation of the fibril.

With the exception of those outliers in CTE fibrils, the residues that held the strongest sway over monomer incorporation across all strains of tau were those that stabilized the protofilament monomer fold, rather than those that held protofilament layers together. To the authors, this suggests that as a fibril grows, the incoming monomer may twist into a particular fold before, or at the same time as, it is being incorporated. This, rather than cross-protofilament contacts, might be the rate-limiting step for fibril formation, they posit.

Remarkably, the researchers found that, similar to cryo-EM-defined folds, the incorporation patterns of alanine mutants classified tau fibrils by neuropathological diagnosis in an unbiased manner (image below). “These results indicated that the seeding activity in human samples encoded information sufficient to discriminate tauopathies based on underlying neuropathological diagnosis,” the authors wrote. “The fidelity of intracellular tau seeding allowed unbiased classification of tauopathies [by this assay].”

That this seeding-based technique aligns so well with the tauopathy family tree traced by cryo-EM structures adds credence to the prion-like properties of tau propagation, Diamond and colleagues believe. What’s more, it implies that relative to costly cryo-EM, this new technique might be a more broadly accessible way to diagnose different tauopathies, Diamond said.

Signature to Structure. Alanine scan signatures show how substitution with alanine residues across tau’s repeat domains influence incorporation into growing fibrils (left). These signatures overlap with amyloid core residues for different fibril structures, and correspond to published, cryo-EM derived models of each disease protofilament (right). [Courtesy of Vaquer-Alicea et al., Science Advances, 2025.]

Cryo-EM buffs Sjors Scheres and Michel Goedert of MRC Laboratory of Molecular Biology, Cambridge, England, commented that the alanine-scanning technique provides an alternative route to classify tauopathy brain samples. “The authors convincingly show that seeded aggregation in their biosensor cells varies with the position of alanine mutations in the tau sequence, in a tau fold-dependent manner,” they wrote. “Thereby, this method holds the exciting potential to be developed into a widely applicable test for the postmortem diagnosis of tauopathies.”

However, Scheres and Goedert also noted that it remains unclear whether the tau fibrils produced within the biosensor cell lines are exactly the same as those plucked from brain samples. Diamond agreed that, particularly for residues outside the core filament fold, the structures in biosensor lines might not match those in the human brain spot-on. What’s most important, Diamond thinks, is that biosensor lines clearly report the structures found in the brain.

His lab is working on refining the alanine-scanning assay to a few key residues that can accurately distinguish between fibrils from neurodegenerative tauopathies. He envisions using this streamlined assay for postmortem neuropathological diagnosis, and possibly for blood screening of tauopathies. Zeroing in on specific amino acids involved in each fibril fold might also help scientists design small molecules to dismantle them, and/or PET ligands to detect them, he suggested. —Jessica Shugart

Comments

1. • Sjors Scheres
     MRC Laboratory of Molecular Biology
   • Michel Goedert
     MRC Laboratory of Molecular Biology
   • Posted: 31 Jan 2025

   This paper by Vaquer-Alicea et al. describes the development of a new method to classify brain homogenates from different tauopathies, without the need for electron cryo-microscopy structure determination. Previously, cryo-EM structures of tau filaments from the brains of individuals with distinct tauopathies revealed that specific tau folds characterize different diseases, leading to a structure-based classification of disease (Shi et al., 2021) and thereby adding a molecular level of neuropathology (Scheres et al., 2023). However, because cryo-EM is expensive, and requires relatively large amounts of sample, cryo-EM structure determination of brain-derived filaments would be unwieldy as part of routine postmortem neuropathology. Although the different tau folds suggest that it should be possible to develop molecules that bind specifically to the different filaments in the future, such molecules are currently not available.

   The method introduced by Vaquer-Alicea et al. provides an alternative route toward a generally applicable technique to classify tauopathy brain samples. The same group had previously introduced tau biosensor cells, HEK293T cells that overexpress truncated constructs of human mutant tau, coupled to fluorescent labels. When seeded with small amounts of tau filaments—from recombinant tau or from tauopathy brain homogenates—the seeded aggregation of tau in these cells can be monitored using fluorescence resonance energy transfer (FRET). In the current paper, the authors show that the seeded tau aggregates in the biosensor cells are indeed amyloid filaments, and they introduce a clever use of alanine scanning that allows for specific readouts when tau filaments with different structures are used as seeds.

   For brain homogenates of 21 individuals with four different tauopathies, the authors convincingly show that seeded aggregation in their biosensor cells varies with the position of alanine mutations in the tau sequence, in a tau fold-dependent manner. Thereby, this method holds the exciting potential to be developed into a widely applicable test for the postmortem diagnosis of tauopathies.

   Because the fluorescent protein labels hampered cryo-EM structure determination, the authors could not determine the structures of the seeded aggregates that form inside the biosensor cells. Previously, we showed that seeded aggregation of overexpressed human tau in SH-S5Y5 cells yielded filaments that resembled, but were not identical, to the filaments that were used as seeds (Tarutani et al., 2023). It is possible that the same is also true for the HEK293T biosensor cells. Therefore, care should be taken in the interpretation of the alanine scan results in terms of the structures of the input seeds. Future cryo-EM structure determination of the seeded aggregates, possibly without their fluorescent labels, would resolve these uncertainties, and would add valuable information to answer the question what determines the formation of specific tau folds in the different diseases.

   References:

   Shi Y, Zhang W, Yang Y, Murzin AG, Falcon B, Kotecha A, van Beers M, Tarutani A, Kametani F, Garringer HJ, Vidal R, Hallinan GI, Lashley T, Saito Y, Murayama S, Yoshida M, Tanaka H, Kakita A, Ikeuchi T, Robinson AC, Mann DM, Kovacs GG, Revesz T, Ghetti B, Hasegawa M, Goedert M, Scheres SH. Structure-based classification of tauopathies. Nature. 2021 Oct;598(7880):359-363. Epub 2021 Sep 29 PubMed.

   Scheres SH, Ryskeldi-Falcon B, Goedert M. Molecular pathology of neurodegenerative diseases by cryo-EM of amyloids. Nature. 2023 Sep;621(7980):701-710. Epub 2023 Sep 27 PubMed.

   Tarutani A, Lövestam S, Zhang X, Kotecha A, Robinson AC, Mann DM, Saito Y, Murayama S, Tomita T, Goedert M, Scheres SH, Hasegawa M. Cryo-EM structures of tau filaments from SH-SY5Y cells seeded with brain extracts from cases of Alzheimer's disease and corticobasal degeneration. FEBS Open Bio. 2023 Aug;13(8):1394-1404. Epub 2023 Jul 7 PubMed.

   View all comments by Sjors Scheres View all comments by Michel Goedert
2. • Lary Walker
     Emory University
   • Posted: 31 Jan 2025

   The strain phenomenon—i.e., the concept that a single proteopathic agent can engender multiple manifestations of disease—has long been an enigmatic piece of the prion puzzle. In microbiology, “strain” classically refers to variant microbes within a given species; whereas microbial strain differences are genetically encoded, evidence has gradually accumulated that the essential feature of prion strains is variation of the 3D structure of the misfolded prion protein. In recent years, this concept has increasingly been applied to the tau and Aβ proteins, which have prion-like functionality and are central to Alzheimer's disease. However, the precise mechanisms driving strain-like variation have been uncertain.

   With the emergence of powerful methods for visualizing the architecture of folded proteins (in particular cryo-electron microscopy), the molecular structural basis of proteopathic strains is coming into focus. Cryo-EM beautifully shows how the amino acids are arranged in amyloids, but it can only furnish limited information on how specific amino acids influence the protean nature of amyloidogenic proteins. Using an ingenious tau "biosensor" assay and alanine substitution scanning, Vaquer-Alicea and colleagues report that different tauopathies (including Alzheimer's) can be differentiated by the pattern of incorporation of systematically mutated (alanine-substituted) tau into intracellular tau aggregates.

   In this paradigm, alanine acts as an inert residue that functionally neutralizes the location of the substituted amino acid, thereby reporting on the role of each amino acid in the interaction of tau monomers with pre-existing tau assemblies within cells. Thus, whereas cryo-EM yields information on the overall structure of proteopathic strains, the alanine substitution/biosensor method highlights the amino acids that most strongly govern strain-like properties. 

   This approach is a welcome addition to the technical armamentarium for defining the molecular basis of proteopathic strains, and it furnishes persuasive new support for the key role of prionic mechanisms in tauopathies. The findings underscore the importance of amino acids in the amyloid core for defining the strain-like variations of tau. They also affirm the influence of the cellular context—the host cells—on the prion-like propagation of tau strains. 

   Discerning the architecture of proteopathic strains is an important step toward understanding the clinical and pathological heterogeneity of Alzheimer's disease and probably many other degenerative disorders. In this regard, the alanine substitution/tau biosensor system might serve as a prototype for the development of new models for interrogating strain-like variations in diverse pathogenic proteins.

   View all comments by Lary Walker

Make a Comment

To make a comment you must login or register.

References

News Citations

1. Flock of New Folds Fills in Tauopathy Family Tree 5 Oct 2021
2. Cellular Biosensor Detects Tau Seeds Long Before They Sprout Pathology 9 Oct 2014
3. Widely Used Tau Seeding Assay Challenged 22 May 2020

Mutations Citations

1. MAPT P301L
2. MAPT V337M

Further Reading

No Available Further Reading