Skip to main content
NCBI home page
ACS AuthorChoice logoLink to ACS AuthorChoice
. 2023 Dec 28;63(2):194–201. doi: 10.1021/acs.biochem.3c00425

Reconstitution of the Alzheimer’s Disease Tau Core Structure from Recombinant Tau297–391 Yields Variable Quaternary Structures as Seen by Negative Stain and Cryo-EM

Calina Glynn †,‡, Joshua E Chun †,‡, Cameron C Donahue †,‡, Monica J S Nadler †,‡, Zhanyun Fan †,‡, Bradley T Hyman †,‡,*
PMCID: PMC10795186  PMID: 38154792

Abstract

graphic file with name bi3c00425_0005.jpg

The protein tau misfolds into disease-specific fibrillar structures in more than 20 neurodegenerative diseases collectively referred to as tauopathies. To understand and prevent disease-specific mechanisms of filament formation, in vitro models for aggregation that robustly yield these different end point structures will be necessary. Here, we used cryo-electron microscopy (cryo-EM) to reconstruct fibril polymorphs taken on by residues 297–391 of tau under conditions previously shown to give rise to the core structure found in Alzheimer’s disease (AD). While we were able to reconstitute the AD tau core fold, the proportion of these paired helical filaments (PHFs) was highly variable, and a majority of filaments were composed of PHFs with an additional identical C-shaped protofilament attached near the PHF interface, termed triple helical filaments (THFs). Since the impact of filament layer quaternary structure on the biological properties of tau and other amyloid filaments is not known, the applications for samples of this morphology are presently uncertain. We further demonstrate the variation in the proportion of PHFs and PHF-like fibrils compared to other morphologies as a function of shaking time and AD polymorph-favoring cofactor concentration. This variation in polymorph abundance, even under identical experimental conditions, highlights the variation that can arise both within a lab and in different laboratory settings when reconstituting specific fibril polymorphs in vitro.

Introduction

A collection of the most common progressive neurodegenerative diseases all share a common feature of accumulation of abnormal fibrillar aggregates of the protein tau (tau_human, UniProtKB entry P10636) in the brain. These diseases are collectively referred to as tauopathies and consist of over 20 neurodegenerative diseases, including Alzheimer’s disease, progressive supranuclear palsy, corticobasal degeneration, and Pick’s disease. The most common tauopathy, Alzheimer’s Disease (AD), ranks in the top ten chronic conditions and top five causes of death among adults aged 65+.1 However, isolation of these tau aggregates from the human brain for further study is inherently messy. These aggregates are known to copurify with lipid vesicles, other proteins, including noteworthy influencers of AD like APOE,2 and fibrillar aggregates of other proteins like amyloid β and TMEM106B,3 particularly in cases with copathology. These components are not readily fully separated from one another, resulting in a range of isolation protocols being used in different laboratories with variable target purity. This raises the possibility that isolation protocol-dependent impurities can play a role in the biological activity of the isolated material.

A mixture of structures, both from the same protein and from additional proteins,3,4 interfere with the ability to draw a direct relationship between a single structure and its contribution to disease-associated properties. Generation of recombinant structures from the same starting material would allow a direct relationship to be drawn between structure and bioactivity without the possibility of disease-specific impurities and alternate protein conformations as influencers. Likewise, understanding the mechanism of fibril formation of specific polymorphs will require the ability to form disease-relevant structures in vitro. Likewise, robust generation of large quantities of specific structures found in disease could aid in structure-based design of aggregation-inhibitory and disaggregating compounds.

However, generating fibrils from full-length recombinant tau has been challenging, and cofactors are typically required. Smaller fragments of tau composed of approximately 100 residues of mixed isoform composition within the microtubule binding repeats were identified as the protease-stable core of paired helical filaments (PHFs) from AD.5,6 This fragment was found to terminate at residue 391 and was hypothesized to either be cleaved from full-length tau by endogenous proteases as an event leading to PHF assembly or cleaved after assembly into a filament precursor structure.7 Soluble 297–391 has been shown to be internalized by cells without being toxic, while fibrils of the same fragment were difficult to internalize but toxic.8 Unlike full-length tau, fibrillization of this fragment—297–391—did not require heparin or other cofactors.9,10 Although ThS curves were similar under both reducing and nonreducing conditions, more abundant filaments are observed by transmission electron microscopy (TEM) under reducing conditions.10 These fibrils have been shown to be structurally consistent with the PHF structure found in AD by ssNMR11 and atomic force microscopy (AFM),12 but whether the structure truly matched the PHF structure was not known.

It has recently been demonstrated that the tau PHF core structure seen in AD13 can be reconstituted in vitro using this same tau fragment isolated from Escherichia coli under specific purification and fibrillization conditions.14 However, amyloid fibril polymorph structures exist in closely positioned energetic minima where favor of one structural arrangement over another may be based on minute alterations in fibrillization conditions. A nonexhaustive list of these variables includes alterations in chemical environment; agitation parameters, including shaking speed, shaking instrument, instrument oscillation radius, sample vessel, and sample volume; solution viscosity; temperature and how it is applied to the sample; ionic strength and size of ions present; protein purification method; product purity; and precise protein construct selected. This makes reproduction of a particular amyloid fibril polymorph tedious, particularly when considering differences in equipment, reagents, and practices between laboratories.

To further complicate robust fibril polymorph generation, it has been seen that structures that cannot be distinguished even from negative stain electron micrographs based on measures of fibril pitch and width can harbor distinct atomic arrangements composed of identical constituent residues.14 In some cases, these fibrils are generated under very similar or identical fibrillization conditions, where only the shaking speed,14 instrument,15 protein concentration,16 reducing agent concentration,16 ion concentration,16 or nothing detectable16 have been changed. For these reasons, it cannot be assumed that a fibril polymorph has been reconstituted in a different lab, with different instrumentation, or using different reagents without verification using an unambiguous high-resolution reconstruction method that allows for visualization of specific residue locations in three-dimensional (3D) space such as cryo-electron microscopy (cryo-EM).

Quantification of in vitro-generated fibril polymorphs from a cryo-EM data set often relies on a single preparation, where preparation-to-preparation variations in relative abundance of different morphologies are not often described. In contrast, patient-to-patient variation in relative polymorph abundance is commonly quantified by cryo-EM,17,18 often with extreme variation of unknown neurological consequence. Here, we aimed to address this commonly overlooked void in knowledge about in vitro polymorph variation using cryo-EM to confirm underlying fibril structures and negative stain EM to quantify variation in polymorph abundance between preparations.

Materials and Methods

Cloning of Untagged Tau 297–391

The gene for an N-terminally 6xHis tagged version of tau 297–391 with a 3C cleavage site following the affinity tag was first synthesized and ligated into a pET28A(+) vector between NcoI and Xho1 restriction sites by Twist Bioscience. The region encoding the 6xHis tag was removed by digestion with NcoI (NEB) and NdeI (NEB). The resulting plasmid was gel-purified, and a duplex oligo (IDT) containing a unique KpnI site was inserted using NEBuilder HiFi DNA Assembly (NEB). Duplex oligo sense sequence: 5′-CTCTAGAAATAATTTTGTTTAACTTTAAGAAGGAGATATAggtaccATGATCAAACACGTCCCGGGAGGCGGCAGTGTGCAAATA-3′. Positive clones were screened by a KpnI (NEB) digest and sequence-verified at the MGH DNA core. This yielded a tag-free gene encoding residues 297–391 of tau only.

Protein Expression and Purification

A pET28a(+) plasmid encoding the sequence of the human tau protein 297–391 without purification tags was transformed into E. coli BL21 (DE3) cells (Thermo Scientific). Cells were grown at 37 °C with shaking at 220–230 rpm until an optical density at 600 nm (OD600) reached approximately 0.6. Protein overexpression was induced by the addition of 100 μg/mL IPTG, followed by 2 h of growth under the same conditions. Cells were harvested, snap-frozen, and stored at −80 °C until protein purification.

The cell pellet was resuspended in 25 mL/L cell culture with 50 mM MES (pH: 6.0), 10 mM EDTA, 10 mM DTT, with 0.1 mM PMSF, and Halt Protease and Phosphatase Inhibitor Cocktail (Thermo Scientific) added fresh. The cells were lysed with two runs of French pressure lysis and centrifuged at 15,000g for 30 min. The supernatant was removed and filtered with a 0.45 μM Syringe filter (Merck Millipore); the pellet was discarded. The roughly 30 mL of filtered supernatant was next injected over a 5 mL HiTrap Capto S cation exchange column (Cytiva) attached to an ÄKTA Pure (Cytiva) using a 50 mL Superloop injector/connector (Cytiva). The sample was injected with 40 mL of wash buffer (50 mM MES (pH: 6.0), 10 mM EDTA, 10 mM DTT, supplemented with 0.1 mM PMSF). Next, the column was washed with 50 mL of wash buffer. The sample was eluted from the column using a linear gradient over 30 mL from wash buffer to an elution buffer consisting of the wash buffer with 1 M NaCl, with 2 mL fractions collected. Fractions with high ultraviolet (UV) signal were run on a 10% Bis-Tris gel (Invitrogen) in MES SDS Running Buffer (Invitrogen), followed by staining with SimplyBlue SafeStain (Invitrogen) (Figure S1A,B).

Fractions containing tau 297–391 were pooled, and ammonium sulfate was added at 0.3 g/mL to precipitate proteins. Following a 30 min incubation while rotating at 4 °C, the sample was centrifuged for 30 min at 15,000g at 4 °C. The supernatant was removed, and the pellet was resuspended in 1–2 mL of 10 mM phosphate bBuffer, pH 7.2.

The resuspended protein was next run over a Superdex 75 Increase, 10/300 GL (Cytiva) size exclusion column attached to a KTA Pure system (Cytiva) using a 5 mL capillary loop. 0.5 mL fractions were collected, starting 9 mL after injection of protein, for 15 mL total. Fractions with high UV signal were run on a 10% Bis-Tris gel (Invitrogen) in MES SDS running buffer (Invitrogen), followed by staining with SimplyBlue SafeStain (Invitrogen). Untruncated protein-containing fractions that did not have any increase in conductivity indicative of remaining salts were pooled and diluted to a final concentration of 6 mg/mL (or 600 μM for this 10.2 kDa fragment), as determined by Pierce BCA Protein Assay Kit (Thermo Scientific) with 10 mM phosphate buffer, pH 7.2 (Figure S1C–E). Protein was flash-frozen for storage at −80 °C until further use.

Fibril Formation

Fibril formation was performed using a FLUOstar Omega instrument in a Corning 96-Well Black Polystyrene Microplate (Thermo Fisher). In an individual well, the purified protein sample was thawed and diluted to 4 mg/mL (or 400 μM) using 10 mM PB, 10 mM DTT, with salts/cofactors added last to a final volume of 100 μL. In a separate well, conditions were replicated with the addition of 25 μM ThT to track the formation of fibrils over time. Readings for fluorescence of ThT were taken every 15 min for 48 h with orbital shaking at 200 rpm and 37 °C. The non-ThT wells would be used for negative stain electron microscopy, with 3 μL of sample pulled at time points from 24 to 48 h.

Negative Stain Electron Microscopy

For negative stain imaging, 3 μL of each fibril sample was applied to a 200 mesh Cu F/C grid and allowed to incubate for 2 min before excess liquid was wicked away, and the grid was washed with 3 μL of water. The grid was then wicked again, and 3 μL of 2% uranyl acetate was applied for 90 s before being wicked and the grid allowed to dry. Images were recorded using either a JEOL 1011 electron microscope at Massachusetts General Hospital or a Thermo Scientific Tecnai 12 electron microscope at the Harvard Medical School Electron Microscopy Suite.

For quantification of fibril polymorphs, ten images from each condition were acquired using a JEOL 1011 electron microscope at 100,000×, and fibrils with PHF-like morphology (10/20 nm min/max width, ∼80 nm crossover distance) were manually counted and compared against the total number of fibrils. For each set of ten images, the total number of fibrils ranged from 54 to 208 for the no-cofactors condition, with the exception of the 24 h time point where no fibrils were found. For all other conditions, 80–386 total fibrils were counted (Table S3). Each experimental condition for 0–400 mM MgCl2 was repeated 3 times on different days, 0.1–0.2 μg/mL dextran sulfate was repeated 2 times, and fibril polymorphs were quantified.

Statistical Analysis

Statistical analyses were carried out using GraphPad Prism 9. To compare time to half-maximum fluorescence of ThT curves, a one-way ANOVA was used, followed by post hoc multiple comparisons between each group where relevant. Data in all graphs are represented as mean and standard deviation from the mean.

Cryo-EM Data Collection

Fibrils were diluted 20× into fibrillization buffer before applying 1.8 μL to each side of a glow discharged R 2/1 300 mesh carbon Cu grid. Grids were plunge frozen into liquid ethane using a Thermo Scientific Vitrobot Mark IV set to 100% humidity with a blot time of 2 s and blot force 10. Cryo-EM images were collected at the Harvard Medical School Cryo-EM facility by using a Thermo Scientific TALOS Arctica equipped with a K3 camera. SerialEM was used to collect 1182 micrographs at a magnification of 36,000× and a pixel size of 1.1 Å with a total dose of 50 e–/Å2 spread over 47 frames and a 4.694 s total exposure with a target defocus from −1.0 to −2.0 Å.

Helical Reconstruction

Patchwise motion correction was carried out using 7 × 5 patches per micrograph, with frame alignment carried out using MotionCorr2 implemented in RELION 4.0.19 Defocus values were determined using Gctf.20 Particles were manually selected in two separate groups, fibrils with dimensions resembling PHFs (63,870 particles) and all other morphologies (200,587 particles) for a calculated 24% of particles resembling PHFs. Particles were first extracted with a 768 pixel box binned to 256 pixels, followed by two-dimensional (2D) classification in RELION 4.0 to remove undesirable particles.

For PHFs, particles were re-extracted using a 256 pixel box, and particles with a CtfMaxResolution poorer than 6 Å were removed before 3D classification using EMD-14063 filtered to 15 Å as a reference. A rise of 2.36 Å and a twist of 179.47 were used to yield an initial reconstruction with an estimated resolution of ∼6 Å before progressing to gold standard 3D autorefinement where the 3D classification map filtered to 10 Å was used as a reference. Masking and postprocessing were carried out in RELION 4.0. The final resolution estimate of 4.7 Å was calculated based on the Fourier shell correlation (FSC) coefficient of 0.143 between independently refined half maps.

For all other morphologies, classes from the first round of 2D classification were used to distinguish between polymorphs and calculate the proportions of the remaining morphologies compared to all particles (including PHFs). Classes with the triple helical filament (THF) morphology were re-extracted using a box size of 1024 pixels binned to 256 pixels in order to capture a full crossover. Classes were used to measure the crossover distance of ∼943 Å and generate an initial 3D model in RELION 4.0. Particles with a CtfMaxResolution better than 5.5 Å were re-extracted using a box size of 256 pixels and underwent 3D classification using the generated initial model filtered to 10 Å as a reference, a twist value of −0.907, and a helical rise of 4.75 Å. 3D autorefinement, masking, and postprocessing were carried out in RELION 4.0, yielding a final resolution of 4.8 Å based on the FSC coefficient at 0.143 criterion described above.

Particles belonging to classes with quadruple helical filament (QHF) morphology were re-extracted with a box size of 1152 pixels binned to 256 pixels in order to capture an entire crossover. One of these classes was used to generate an initial 3D model, where a rough outline of four C-shaped protofilaments can be seen, but due to low particle numbers (4891), further reconstruction in 3D was unsuccessful.

All remaining classes that were not THFs or QHFs were of untwisted morphologies, whose structures were not determined here.

Results

We first sought to generate the PHF core structure found in AD. Starting from a fibrillization condition containing 200 mM MgCl2, we were able to generate fibrils composed of the same C-shaped building blocks found in AD PHFs. However, we obtained a mixture of quaternary structures where 15–26% (n = 3 independent fibrillization experiments) of total fibrils harbored PHF-like morphology by negative stain EM after following the 48 h aggregation protocol described to form 96% PHF structures confirmed by cryo-EM14 or 8% PHF-like structures seen by negative stain electron microscopy (EM) in another study16 (Table S1).

Impact of Time Spent Shaking on PHF-Like Fibril Abundance

Monitoring aggregation using the amyloid-specific dye thioflavin T (ThT) revealed that peak fluorescence was achieved within ∼30 h of reaction initiation with some variation (Figure 1A). Thus, we sought to assess the relative abundance of PHF-like fibrils compared to all fibrils formed over time in the event that PHFs formed sooner than other morphologies or coalesced to form higher-order structures at later time points. We found that PHF-like fibrils formed early at varying relative abundance, with 8–57% (n = 3) of fibrils displaying PHF-like morphology after 24 h of shaking for different replicates by negative stain EM (Figure 1B,1C). This fraction relative to total fibrils did not change in a statistically significant way as time spent shaking increased.

Figure 1.

Figure 1

Open in a new tab

Fibrillization of Tau 297–391 in the presence of 200 mM MgCl2. (A) Thioflavin T (ThT) was used to track fibril formation over time, with peak fluorescence reached within 24–30 h of shaking (n = 5). (B) The fraction of fibrils with PHF-like dimensions (10–20 nm width) and pitch (∼75–80 nm crossover distance) were manually counted from 10 micrographs per time point (n = 3 replicate wells per time point). Total number of fibrils counted per point ranged from 124 to 386. Bar represents the mean, with error bars representing the standard deviation. Each shape represents fibrils taken from the same well at different time points. (C) Representative images from each time point with PHF-like structures marked by arrows. Scale bars: 200 nm.

To determine whether the fibrils generated harbored the PHF core structure, we next turned to cryo-EM. Due to a lower abundance of total fibrils at 24 and 27 h time points, fibrils that had been shaking for 30 h were used for cryo-EM studies. These were found to be mainly triple helical filaments (THFs, 65%), PHFs (24%), and a small number of quadruple helical filaments (QHFs, 2%) with a minor population (9%) of untwisted or irregularly twisting species whose three-dimensional structures were not determined here (Figure 2 and Table S2).

Figure 2.

Figure 2

Open in a new tab

Cryo-EM reconstruction of PHFs, THFs, and QHFs. (A) Negative stain micrograph of fibrils formed with 200 mM MgCl2 at 30 h of shaking, highlighting the most abundant species; paired helical filaments (PHFs) and triple helical filaments (THFs). (B) Zoomed in and enhanced contrast for boxed regions in panel (A). (C) Frozen-hydrated fibrils, with PHFs marked by arrows and THFs marked by asterisks. (D) 2D classification and cross section from PHFs. Large 2D classification box size = 768 pixels/845 Å, smaller 2D classification box size = 384 pixels/422 Å, 3D refinement box size = 256 pixels/282 Å. (E) Same as panel (D) but for THFs where large 2D classification box size = 1024 pixel/1126 Å. Smaller 2D classification and 3D refinement box sizes are the same as described in panel (D). (F) Same as panel (D) but for quadruple helical filaments (QHFs) where large box size = 1152 pixel/1267 Å and smaller 2D classification box size the same as described in panel (D). Cross section shown from the initial model generated in RELION.

Impact of Magnesium Chloride Concentration on PHF-Like Fibril Abundance Over Time

Since the addition of 200 mM MgCl2 has been used to favor PHFs over THFs and QHFs,14 we sought to quantify the relationship between MgCl2 concentration and PHF-like morphology abundance. Without MgCl2, we were unable to find fibrils after 24 h of incubation. After 30 and 48 h of shaking, fibrils had formed, but <1% of total fibrils harbored PHF-like morphology by negative stain EM for both time points (Figure 3). This illustrates that MgCl2 is helpful in favoring PHFs over other structures but does not guarantee that all or even most structures formed will be PHFs regardless of time spent shaking.

Figure 3.

Figure 3

Open in a new tab

Impact of MgCl2 concentration on fraction of PHF-like fibrils. (A) ThT fluorescence curves for fibrils formed with 0 mM (pink), 200 mM (purple), and 400 mM (blue) MgCl2 (n = 5). 200 mM curves presented in Figure 1A are shown again here for ease of comparison. (B) Time to half-maximum fluorescence as a function of MgCl2 concentration (n = 5). Fibrils formed statistically significantly faster (Tukey’s multiple comparison test, P = 0.0117) in 400 mM MgCl2 compared to 0 mM MgCl2. Error bars represent one standard deviation. (C) Percent PHF-like fibrils identified by negative stain EM as a function of time for each concentration of MgCl2 used. For 200 mM MgCl2, data points for 24, 30, and 48 h from Figure 1B are shown again here for ease of comparison. Bar represents the mean, with error bars representing the standard deviation. (D) Representative negative stain electron micrographs from fibrils formed at each MgCl2 concentration at each time point (n = 10 micrographs per data point shown in panel (C)). Where PHF-like fibrils were found (arrows), a boxed region was enlarged to show PHF-like morphology. Scale bar = 200 nm.

In another work, doubling the MgCl2 concentration from 200 mM to 400 mM increased the fraction of PHF-like fibrils by negative stain EM from 8 to 30%.16 When we increased MgCl2 to 400 mM, we observed faster and more consistent aggregation based on the time to half the maximum fluorescence of ThT than when no MgCl2 was included (Figure 3C). Despite this change in aggregation kinetics, the relative abundance of PHF-like fibrils did not change significantly from 14% (range 9–23%) at 200 mM MgCl2 to 15% (range 12–21%) at 400 mM MgCl2 after 30 h of shaking. Similarly, neither a statistically significant change nor a visually apparent trend was observed in the proportion of PHF-like structures between 200 and 400 mM MgCl2 from 24 to 48 h.

Dextran Sulfate and Other Conditions Reported to Form PHFs

Since other conditions were also reported to form PHF structures from this same fragment in vitro,14,16 we assessed these conditions for the formation of PHF-like structures as well. Similar to 200 mM MgCl2, addition of 0.1 μg/mL dextran sulfate also favored the formation of PHF-like structures, although with a similar degree of structural impurity and prep-to-prep variation (Figure 4). In two replicates, 17 and 35% of total fibrils favored a PHF-like morphology by negative stain EM. For two replicates containing dextran sulfate doubled to 0.2 μg/mL, 27 and 69% of total fibrils were PHF-like after 24 h. After 48 h of incubation, less variation was observed between preparations, with 5 and 14% of fibrils formed in the presence of 0.1 μg/mL dextran sulfate displaying PHF-like morphology. With 0.2 μg/mL dextran sulfate, a similarly low fraction of PHF-like fibrils was found for both preps after 48 h (4 and 8%).

Figure 4.

Figure 4

Open in a new tab

Impact of dextran sulfate concentration on fraction of PHF-like fibrils. (A) ThT curves for fibrils formed in the presence of 0.1 (lime) or 0.2 (teal) μg/mL dextran sulfate. (B) Time to half-maximum fluorescence as a function of dextran sulfate concentration, where error bars represent one standard deviation. (C) Fraction PHF-like fibrils compared to total fibrils as a function of dextran sulfate concentration over time. Bar represents the mean, with error bars representing the standard deviation. (D) Representative micrographs (n = 10 micrographs per data point in panel (C)) for each condition, with a maximum of five fibrils with PHF-like morphology highlighted (arrows) and one boxed PHF-like fibril enlarged. Scale bar = 200 nm.

This illustrates that under conditions containing dextran sulfate, early time points (24 and 30 h) are highly variable in terms of fraction of total fibrils that have PHF-like morphology, while later time points (48 h) consistently contain a smaller fraction of PHF-like fibrils compared to total fibrils in our hands.

Discussion

In AD, tau filaments are found as two ultrastructures that each contain two identical C-shaped protofilament building blocks per layer, termed paired helical filaments and straight filaments. Fibrils composed of one, three, or four C-shaped protofilaments per layer have not been found in AD. However, fibrils with a single C-shaped protofilament per layer can be found upon seeding human neuroblastoma SH-SY5Y cells expressing 1N3R tau with AD-derived tau fibrils.21 Tau filaments composed of varying numbers of these same C-shaped building blocks can be readily formed from truncated tau under a variety of conditions14,16 including here, but generating the AD tau ultrastructure where exactly two C-shaped building blocks per layer come together in a specific arrangement robustly is still a challenge.

Taken together, our work demonstrates that variation in polymorph relative abundance can exist both between laboratories and between experiments performed in the same lab. For our cryo-EM study using 200 mM MgCl2 to spur fibrillization for 30 h, 24% of particles could be classified as PHFs and >92% of particles fell into classes with the C-shaped building blocks found in AD. However, across replicates of the same condition sampling from 24 to 48 h, we were able to find anywhere from 8 to 57% PHF or PHF-like fibril populations depending on the time point and replicate. These same conditions were used by other groups to generate preparations with 8% of PHF-like fibrils by negative stain EM16 or 96% of PHF particles by cryo-EM.14 We note that earlier time points seemed to have the most variation while later time points were more consistent in terms of fraction PHFs in our study. This change over time and across replicates was also observed in other groups, where 50–90% of fibrils took on the PHF structure after 12 h, depending on the replicate.22

For our other conditions outside of that utilizing 200 mM MgCl2 after 30 h of fibril formation, we note that fibrils are described as “PHF-like” since we have not determined these structures using cryo-EM, and it is possible that other structures resembling PHFs have been formed instead.

It remains to be seen if the differences in quaternary structure we observed (PHF vs THF) have an impact on the biological properties of these filaments, but pure populations of each ultrastructure will be critical to answer these questions. While one of our replicates had a high percentage of PHF-like structures after 24 h, fibril formation was incomplete based on the difficulty of finding fibrils compared to later time points. This indicates that a large portion of the solution remained nonfibrillar at that time. This adds an additional variable to account for in biological experiments using material with mixed populations, including fibril polymorphs, monomers, and oligomers, at ill-defined individual concentrations. Furthermore, variation in aggregation kinetics and relative quantities of different fibril populations (Figure 1) were observed, which would require quantification of different populations at each time point in a well-dependent manner each time the protein is fibrillized, which is not ideal.

In sum, we replicate the ability to form PHF fibril core structures using only the truncated 297–391 tau peptide yet also show that the efficiency of forming the biologically relevant 3D structure varies by both predictable and unknown experimental conditions. This supports the paradigm that different conformational polymorphs exist in closely positioned energetic minima. Nonetheless, within human brain in each of the tauopathies, reproducible forms of the same fibrillar structure occur even despite variations in fibril-associated post-translational modifications and tau isoform composition.23 Understanding the cofactors that contribute to this robust structural homogeneity in patients may provide clues to both understand the forces that drive tau aggregation differentially in the different tauopathies as well as provide possible insight into therapeutic approaches.

Acknowledgments

The authors thank Dr. Richard Walsh (HMS) for assistance with collecting cryo-EM data. The authors would additionally like to thank Dr. Sofia Lövestam, Dr. Michel Goedert, and Dr. Sjors Scheres at the MRC-LMB for advice regarding production of paired helical filaments. This work was supported by the Cure Alzheimer’s Fund, the Massachusetts Alzheimer Disease Research Center (P30AG06241), RF1 AG058674; RF1 AG059789, the JPB Foundation, and R56 AG061196 (BTH), which supported our use of instruments at the Harvard Medical School Molecular Electron Microscopy Suite (MEMS) and cryo-EM facility. C.G. was supported by the NIH T32 Postdoctoral Program to Enrich Translation and Multimodal Research in Alzheimer’s Disease and Related Dementias (T32-AG066592) and a grant from the Toffler Foundation.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.biochem.3c00425.

  • Purification traces and SDS-PAGE gels; FSC curves and data collection and refinement statistics for PHF and THF reconstructions; table with differences in pipeline from protein expression through fibrilization conditions used in this work; Lovestam et al., and Li et al., and table with raw counts of PHF-like fibrils formed under different conditions seen by negative stain EM (PDF)

Author Present Address

§ Structural Biology, Rosalind Franklin Institute, Harwell Science and Innovation Campus, Didcot, OX11 0QS, United Kingdom

Author Contributions

C.G. and B.T.H conceptualized the project. M.J.S.N. and Z.F. generated plasmids. C.C.D. and C.G. produced and purified recombinant proteins. C.G. and J.E.C. performed protein fibrillization experiments and negative stain electron microscopy. J.E.C. performed polymorph quantification from negative stain electron microscopy images. C.G. carried out cryo-EM data collection and fibril reconstruction. C.G., J.E.C., C.C.D., M.J.S.N., Z.F., and B.T.H critically analyzed and provided feedback on the data. C.G. and B.T.H. wrote the manuscript with input from all authors.

The authors declare no competing financial interest.

Notes

Protein Accession IDs Human microtubule-associated protein tau: UniProtKB P10636.

Supplementary Material

bi3c00425_si_001.pdf (400.5KB, pdf)

References

  1. Kochanek K. D.Mortality in the United States, 2020. [PubMed]
  2. Strittmatter W. J.; Saunders A. M.; Goedert M.; Weisgraber K. H.; Dong L. M.; Jakes R.; Huang D. Y.; Pericak-Vance M.; Schmechel D.; Roses A. D. Isoform-specific interactions of apolipoprotein E with microtubule-associated protein tau: implications for Alzheimer disease. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 11183–11186. 10.1073/pnas.91.23.11183. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Chang A.; Xiang X.; Wang J.; Lee C.; Arakhamia T.; Simjanoska M.; Wang C.; Carlomagno Y.; Zhang G.; Dhingra S.; et al. Homotypic fibrillization of TMEM106B across diverse neurodegenerative diseases. Cell 2022, 185, 1346–1355.e15. 10.1016/j.cell.2022.02.026. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Yang Y.; Shi Y.; Schweighauser M.; Zhang X.; Kotecha A.; Murzin A. G.; Garringer H. J.; Cullinane P. W.; Saito Y.; Foroud T.; et al. Structures of α-synuclein filaments from human brains with Lewy pathology. Nature 2022, 610, 791–795. 10.1038/s41586-022-05319-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Wischik C. M.; Novak M.; Thøgersen H. C.; Edwards P. C.; Runswick M. J.; Jakes R.; Walker J. E.; Milstein C.; Roth M.; Klug A. Isolation of a fragment of tau derived from the core of the paired helical filament of Alzheimer disease. Proc. Natl. Acad. Sci. U.S.A. 1988, 85, 4506–4510. 10.1073/pnas.85.12.4506. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Jakes R.; Novak M.; Davison M.; Wischik C. M. Identification of 3- and 4-repeat tau isoforms within the PHF in Alzheimer’s disease. EMBO J. 1991, 10, 2725–2729. 10.1002/j.1460-2075.1991.tb07820.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Novak M.; Kabat J.; Wischik C. M. Molecular characterization of the minimal protease resistant tau unit of the Alzheimer’s disease paired helical filament. EMBO J. 1993, 12, 365–370. 10.1002/j.1460-2075.1993.tb05665.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Pollack S. J.; Trigg J.; Khanom T.; Biasetti L.; Marshall K. E.; Al-Hilaly Y. K.; Rickard J. E.; Harrington C. R.; Wischik C. M.; Serpell L. C. Paired Helical Filament-Forming Region of Tau (297–391) Influences Endogenous Tau Protein and Accumulates in Acidic Compartments in Human Neuronal Cells. J. Mol. Biol. 2020, 432, 4891–4907. 10.1016/j.jmb.2020.05.027. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. del C Alonso A.; Zaidi T.; Novak M.; Grundke-Iqbal I.; Iqbal K. Hyperphosphorylation induces self-assembly of τ into tangles of paired helical filaments/straight filaments. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 6923–6928. 10.1073/pnas.121119298. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Al-Hilaly Y. K.; Pollack S. J.; Vadukul D. M.; Citossi F.; Rickard J. E.; Simpson M.; Storey J. M. D.; Harrington C. R.; Wischik C. M.; Serpell L. C. Alzheimer’s Disease-like Paired Helical Filament Assembly from Truncated Tau Protein Is Independent of Disulfide Crosslinking. J. Mol. Biol. 2017, 429, 3650–3665. 10.1016/j.jmb.2017.09.007. [DOI] [PubMed] [Google Scholar]
  11. Al-Hilaly Y. K.; Hurt C.; Rickard J. E.; Harrington C. R.; Storey J. M. D.; Wischik C. M.; Serpell L. C.; Siemer A. B. Solid-state NMR of paired helical filaments formed by the core tau fragment tau(297–391). Front. Neurosci. 2022, 16, 988074 10.3389/fnins.2022.988074. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Lutter L.; Al-Hilaly Y. K.; Serpell C. J.; Tuite M. F.; Wischik C. M.; Serpell L. C.; Xue W.-F. Structural Identification of Individual Helical Amyloid Filaments by Integration of Cryo-Electron Microscopy-Derived Maps in Comparative Morphometric Atomic Force Microscopy Image Analysis. J. Mol. Biol. 2022, 434, 167466 10.1016/j.jmb.2022.167466. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Fitzpatrick A. W. P.; Falcon B.; He S.; Murzin A. G.; Murshudov G.; Garringer H. J.; Crowther R. A.; Ghetti B.; Goedert M.; Scheres S. H. W. Cryo-EM structures of tau filaments from Alzheimer’s disease. Nature 2017, 547, 185–190. 10.1038/nature23002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Lövestam S.; Koh F. A.; van Knippenberg B.; Kotecha A.; Murzin A. G.; Goedert M.; Scheres S. H. Assembly of recombinant tau into filaments identical to those of Alzheimer’s disease and chronic traumatic encephalopathy. eLife 2022, 11, e76494 10.7554/eLife.76494. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Al-Hilaly Y. K.; Foster B. E.; Biasetti L.; Lutter L.; Pollack S. J.; Rickard J. E.; Storey J. M. D.; Harrington C. R.; Xue W.; Wischik C. M.; et al. Tau (297–391) forms filaments that structurally mimic the core of paired helical filaments in Alzheimer’s disease brain. FEBS Lett. 2020, 594, 944–950. 10.1002/1873-3468.13675. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Li X.; Zhang S.; Liu Z.; Tao Y.; Xia W.; Sun Y.; Liu C.; Le W.; Sun B.; Li D. Subtle change of fibrillation condition leads to substantial alteration of recombinant Tau fibril structure. iScience 2022, 25, 105645 10.1016/j.isci.2022.105645. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Falcon B.; Zhang W.; Schweighauser M.; Murzin A. G.; Vidal R.; Garringer H. J.; Ghetti B.; Scheres S. H. W.; Goedert M. Tau filaments from multiple cases of sporadic and inherited Alzheimer’s disease adopt a common fold. Acta Neuropathol. 2018, 136, 699–708. 10.1007/s00401-018-1914-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Falcon B.; Zhang W.; Murzin A. G.; Murshudov G.; Garringer H. J.; Vidal R.; Crowther R. A.; Ghetti B.; Scheres S. H. W.; Goedert M. Structures of filaments from Pick’s disease reveal a novel tau protein fold. Nature 2018, 561, 137–140. 10.1038/s41586-018-0454-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Kimanius D.; Dong L.; Sharov G.; Nakane T.; Scheres S. H. W. New tools for automated cryo-EM single-particle analysis in RELION-4.0. Biochem. J. 2021, 478, 4169–4185. 10.1042/BCJ20210708. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Zhang K. Gctf: Real-time CTF determination and correction. J. Struct Biol. 2016, 193, 1–12. 10.1016/j.jsb.2015.11.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Tarutani A.; Lövestam S.; Zhang X.; Kotecha A.; Robinson A. C.; Mann D. M. A.; Saito Y.; Murayama S.; Tomita T.; Goedert M.. et al. Cryo-EM structures of tau filaments from SH-SY5Y cells seeded with brain extracts from cases of Alzheimer’s disease and corticobasal degeneration Preprint at bioRxiv 2023 10.1101/2023.04.13.536725. [DOI] [PMC free article] [PubMed]
  22. Lövestam S.; Li D.; Wagstaff J. L.; Kotecha A.; Kimanius D.; McLaughlin S. H.; Murzin A. G.; Freund S. M. V.; Goedert M.; Scheres S. H. W.. Disease-specific tau filaments assemble via polymorphic intermediates Preprint at bioRxiv 2023 10.1101/2023.07.24.550295. [DOI] [PMC free article] [PubMed]
  23. Wesseling H.; Mair W.; Kumar M.; Schlaffner C. N.; Tang S.; Beerepoot P.; Fatou B.; Guise A. J.; Cheng L.; Takeda S.; et al. Tau PTM Profiles Identify Patient Heterogeneity and Stages of Alzheimer’s Disease. Cell 2020, 183, 1699–1713.e13. 10.1016/j.cell.2020.10.029. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

bi3c00425_si_001.pdf (400.5KB, pdf)

Articles from Biochemistry are provided here courtesy of American Chemical Society

RESOURCES