The cDNA encoding the Tau isoform D (383aa) has been cloned into the BspEI site of the plasmid pECFP-EYFP already available in the lab. Both the forward and reverse cloning primers contain the RSIVT linker sequence between the BspEI site and the Tau sequence (forward primer: 5′-GTCGTT TCCGGAAGATCTATTGTCACTATGGCTGAG-3′; reverse primer: 5′-AACGACTCCGGA AGTGACAATA- GATCTCAAACCCTG-3′). The monolabeled constructs pECFP-Tau and pTau-EYFP have been generated by subcloning the Tau cDNA into the BspEI site of plasmid pECFP-C1 and pEYFP-N1 (Clontech Laboratories, Inc., Saint-Germain-en-Laye, France) (FWD-BspEI-TAU: 5′-AATTATT- CCGGAATGGCTGAGCCCACGCCAG-3′; REV-BspEI-TAU: 5′-ACTTGATCCGGACAAACCCTGCTTAGGCCAG-3′; FWD- BspEI-RSIVT-TAU: 5′-AATTATTCCGGAAGATCTATTGT CACTAT GGCTGAGCCCACGCCAG-3′; REV-BspEI-RSIVT-TAU:5′-ACTTGAT CCGGAAGTGACAATAGATCTC AAA CCCTGCTTAGGCCAG-3′). The CST mutants have been generated by the Q5 Site-Directed Mutagenesis Kit (New England BioLabs) (CST-AT8mut primers to introduce the mutations S199A-S202A-T205A: Fwd 5′-CCCA GGCGCACCCGGCAGCCGCTCCCGC-3′; Rev5′-GCG CCGGGGGCGCTGTAGCCGCTGCGATCCCC-3′;CST-P301L primers: Fwd 5′-AAACACGTCCTGGGAGGCGGC-3′; Rev 5′-GATATTATCCTTTGAGCCACACTTGGACTG-3′; CST-∆K280 primers: Fwd 5′-AAGCTGGATCTTAGCAAC-3′; Rev 5′-ATTAATTATCTGCACCTTCC-3′). The pTagRFP-tubulin plasmid was purchased from Evrogen (FP145).

HeLa cells and immortalized hippocampal neurons HT22 were maintained in DMEM (GIBCO) supplemented with 10% FCS. The day before the experiment cells were seeded at 10 × 10⁴ cells per well in six-well plates or in Willco dishes (Willcowells). The lipofection was carried out with Effectene (QIAGEN) according to manufacturer’s instructions. Cells have been treated with 1 μM Nocodazole (Sigma) for 30 min, or with 1 μM Paclitaxel (SIGMA) for 10 min. Tau seeds have been produced by in vitro oligomerization. The oligomerization reaction has been performed with 4 mg/ml truncated Tau (297-391)4R in PBS pH 7.2 and 100 μM Heparin (Sigma, H3149) and incubated for 4 days at room temperature. The seeds were purified by centrifugation at 100,000× g for 1.5 h, the pellet was rinsed with PBS. Microtubules labeling in live cells has been obtained by SiR-tubulin Kit (excitation wavelength λ = 633 nm) (SPIROCHROME) in FRET experiments to avoid interference in FRET signal.

For Western blot cells extracts were prepared in lysis buffer supplemented with protease and phosphatase inhibitors by lysis on ice for 30 min. Total proteins were separated by 10% or 8% SDS-PAGE and electro-blotted onto nitrocellulose membranes Hybond-C-Extra (Amersham Biosciences). Membranes were blocked with 5% skimmed milk powder in TBS containing 0.1% Tween 20. For the immunofluorescence cells were fixed with ice-cold 100% methanol × for 5 min. After permeabilization with PBS 1× containing 0.1% Triton-X100 for 10 min, samples were blocked for 1 h with 1% (wt/vol) BSA at RT. The slides were incubated with the primary antibody 1 h at room temperature and with secondary antibodies fluorophore-conjugated 1 h at room temperature. Slides were mounted with Vectashield mounting medium (Vector Laboratories). Primary antibodies were: mouse monoclonal anti-tau (Tau5) ab80579 (abcam); mouse monoclonal anti-a-Tubulin Clone B-5-1-2 (SIGMA-ALDRICH); mouse monoclonal anti-GAPDH (Fitzgerald); mouse monoclonal anti-Tau DC25 (from Novak M.); mouse monoclonal anti-Tau Clone AT8 (MN1020, Thermo Scientific). Secondary antibodies for Western blot analysis were HRP-conjugated anti-mouse or anti-rabbit, purchased from Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA. Secondary antibodies for IF: Alexa Fluor 633 rabbit anti-mouse IgG (Life Technologies).

Images were acquired with the TCS SL laser-scanning confocal microscope (Leica Microsystems, Milan, Italy) equipped with galvanometric stage using a 63×/1.4 NA HCX PL APO oil immersion objective. A heated and humidified chamber mounted on the stage of the microscope was used for live imaging experiments in order to maintain a controlled temperature (37°C) and CO₂ (5%) during image acquisition. An Argon laser was used for ECFP (λ = 458 nm) and EYFP (λ = 514 nm), a Gre-Ne laser for RFP (λ = 543 nm) and a He-Ne laser for λ = 633 nm. For the quantification of morphological parameters such as the total filament length and the number of crossover points the filament tracer option of the IMARIS Bitplane software has been exploited. In detail, these two parameters are deduced by a software plugin that, based on connectivity and fluorescence intensity, automatically detects and segments filamentous structures revealing information about the topology of filaments as the sum of the lengths of all lines and the number of crossover points within the filament.

For sensitized emission FRET experiments, each image was recorded in a spectral mode, by selecting specific regions of the emission spectrum. The donor ECFP was excited at 458 nm and its fluorescence emission was collected between 470 nm and 500 nm (donor channel) and between 530 nm and 600 nm (FRET channel). The acceptor EYFP was excited at 514 nm and its fluorescence emission was collected between 530 nm and 600 nm (acceptor channel). The donor and acceptor fluorophores were excited sequentially. The ImageJ software was used for images analysis and FRET quantification. Briefly, FRET images were corrected from cross-talk between donor and acceptor channel using Youvan’s method (Youvan et al., 1997): F_index = I_FRET−A × I_D−B × I_A, where I_FRET, I_D and I_A are the images of the sample in the FRET, donor and acceptor channel, respectively, after background subtraction and A and B are the fraction of the donor and acceptor leak-through into the FRET channel, respectively. We determined the A parameter in cells expressing only the donor (pECFP plasmid) by acquiring images in donor and FRET channels. A plot of the intensity of each pixel of the FRET channel image as a function of the intensity of the same pixel in the donor channel image (regression graph) was automatically generated by the ImageJ plugin (FRET and colocalization Analyzer) and then fitted with a linear equation from which A parameter was derived. Images displaying saturated pixel has been discarded. For the evaluation of the B parameter cells transiently expressing only the acceptor (pEYFP plasmid) have been acquired in FRET and acceptor channel and then subjected to same process of A in the ImageJ plugin. Typical values in our experimental conditions are A = 0.1 and B = 0.25, respectively. Normalized FRET (NFRET) was performed with the ImageJ software plugin “pixFRET” (Feige et al., 2005) by using: NFRET = F_index/√(I_D × I_A) (Xia and Liu, 2001). Mathematically, NFRET values should range between 0–1; the plugin automatically multiplies these values ×100; NFRET intensities images were represented in false-color obtained by using the “fire” lookup table option in ImageJ software and shown from 0 to 60. We setup this color scale in order to better appreciate also lower NFRET signals. Due to the mathematical subtractions in F_index calculation, NFRET negative values in some pixels are considered as zero. In FRET experiments investigating intramolecular and inter-molecular FRET, to obtain a comparable amount of Tau we transfected cells with 300 ng of CST or 150 ng ECFP-Tau + 150 ng EYFP-Tau (ML/2); to obtain a comparable amount of fluorophores we transfected cells with 300 ng ECFP-Tau + 300 ng EYFP-Tau (ML). Since in a transient transfection the expression of the transgene varies from cell to cell, to perform a reliable NFRET quantification in ML samples we selected cells that express comparable amount of fluorescence intensity in the donor and acceptor channel with respect of CST; to perform the NFRET quantification in ML/2 samples we selected cells that express about half the amount of fluorescence intensity in the donor and acceptor channel with respect of CST. We measured the level of expression by fluorescence intensity in individual cells, under the microscope. FRAP experiments were performed by using the FRAP module coupled to the confocal microscope and consists of three different phases: (1) a pre-bleach phase, in which 10 frames of 512 * 512 pixel images at 1000 Hz have been recorded in order to define the initial level of fluorescence intensity; (2) a photobleaching phase, in which a selected circular ROI with a radius of 2 μm in the cytoplasm of the cell was excited at higher laser power (50% for EYFP) for five frames at 1000 Hz; (3) a post-bleaching phase, in which 120 images have been recorded in order to follow the recovery of the fluorescence intensity in the selected ROI. Fluorescence recovery was extracted from images of the bleached ROI and subjected to the following manipulation steps: (1) background subtraction; (2) first normalization to the initial pre-bleach value of fluorescence intensity; (3) correction for the fluorescence loss: in particular, we evaluated the fluorescence of the whole cell for each time point, before and after the photobleaching phase; then we multiplied every element of the data of the bleached ROI by (F_{pre-wholecell}/F_{post-wholecell}), where F_{pre-wholecell} means prebleach intensity within the whole cell, F_{post-wholecell} stands for postbleach intensity at any given time point. In any case, we set up the experimental parameters of acquisition in order to avoid strong photobleaching (5%–10%); (4) additional normalization to set the first post-bleach point to zero. At least 30 separate FRAP experiments for each sample has been performed. FRAP recovery curves have been fitted by a two phase exponential association function (OriginLab): y = y₀ + A₁^* × (1 − exp(−t/τ₁)) + A₂^* × (1 − exp(−t/τ₂)); fractions A₁ and A₂ reported in the Supplementary Table S1 are defined as: A₁ = A₁^*/(A₁^* + A₂^*) and A₂ = A₂^*/(A₁^* + A₂^*); the mobile fraction is defined as Mob_calc = A₁^* + A₂^*.

To investigate whether, in physiological conditions, the binding to MTs induces Tau protein folding rearrangements, we exploited a conformational sensitive fluorescent sensor by fusing ECFP at the N-terminus and EYFP at the C-terminus of the full length human Tau-D sequence (0N4R) (Figure 1A). It is expected that, depending on the conformation of Tau, a FRET signal could be generated by the decreased distance between the N- and C-termini of Tau molecules, thus realizing a context-dependent Conformational-Sensitive Tau sensor (CST). Mono-labeled fluorescent Tau constructs have been used as controls.

Confocal microscopy imaging in live cells showed that CST decorated the cellular MTs network and colocalized with the tubulin-RFP, both in HeLa cells, that do not express endogenous Tau, and in immortalized hippocampal neurons HT22 (Figure 1B), demonstrating that, under physiological conditions, the CST preserves Tau ability to interact with MTs.

Current views indicate that misfolded Tau drives its own aggregation and spreading in Tauopathy disease progression (Frost et al., 2009; Kfoury et al., 2012; Le et al., 2012; Holmes and Diamond, 2014; Holmes et al., 2014; Mirbaha et al., 2015).

To verify whether CST could detect the seeding induced by molecules potentially involved in pathological self-aggregation, reporter cells expressing CST were exposed to extracellular synthetic Tau seeds, prepared from (297-391)4R recombinant Tau fragment (Kontsekova et al., 2014). Forty-eight hours later the Tau network decorating MTs almost disappeared and concomitantly fluorescent cytoplasmic inclusions appeared (Figure 1C), while the MT network was not affected.

Altogether, these results indicated that CST is sensitive to Tau seeding activities and it allows detecting the Tau displacement from MTs and its aggregation.

To exploit the CST as a conformational probe, we used quantitative sensitized emission FRET microscopy for the detection of protein conformational changes. First, we set FRET parameters in cells expressing the donor and the acceptor fused into a chimeric construct (ECFP-EYFP), as a FRET positive control, and in cells expressing the donor and the acceptor from separate plasmids (pECFP and pEYFP), as a FRET negative control (Supplementary Figure S1). Figure 2A shows donor and acceptor imaging of CST and also a NFRET image (Feige et al., 2005), indicating that CST displayed a FRET-positive signal mostly on MT lattice. Indeed, along selected lines (white line Figure 2B) at the intersection with MTs, the CST displayed NFRET values in the range of 15–30.

To verify whether the observed FRET signal was due to intramolecular or to intermolecular interactions between the Tau N- and C-termini, cells co-expressing monolabeled Tau constructs (ECFP-Tau and Tau-EYFP) have been analyzed. Due to the presence of two fluorophores per Tau molecule in the CST construct, we compared the FRET signals obtained from CST expressing cells, with that obtained from cells expressing monolabeled Tau. We considered either cells expressing the same levels of monolabeled Tau, with respect to CST cells (hence half amounts of fluorophores, ML/2; Figures 2C,D) or cells expressing double levels of monolabeled Tau, with respect to CST cell (hence same amount of fluorophores, ML; Figures 2E,F). In both conditions, the NFRET signal from the cells expressing monolabeled Tau was significantly lower than that observed with CST (Figures 2D,F). The comparative quantification of the FRET values shows a statistically significant difference between the various conditions (Figure 2G).

These results demonstrate that MT-bound CST assumes a hairpin three-dimensional conformation, with its N-terminal and C-terminal domains brought in proximity, allowing FRET to occur. On the contrary, in conditions where only intermolecular FRET would occur, such as when monolabeled Tau fluorescent pairs are expressed, the FRET signal is significantly lower. The FRET signal shows a precise co-localization with the MT network, as revealed by incorporation of fluorescent tubulin (Sir-tubulin) in cellular MTs (Figure 2H), thus demonstrating that the FRET signal arose indeed from a conformation of the CST upon binding to MTs.

To analyze the conformation of soluble Tau monomers, not bound to MTs, CST reporter cells were treated with the MTs disrupting drug Nocodazole (Noc). In this condition, the CST diffused into the cytoplasm (Figure 3A) and concomitantly, the FRET signal was lost (Figures 3A,B), indicating that its N- and C-termini are not close one to another, in this condition. It is noteworthy that the CST rendered soluble by depolymerizing MTs does not show signs of aggregation. Western blot analysis showed that, under these conditions, most of the CST molecules are intact (Supplementary Figure S2A), thus the lack of FRET signal is not due to the cleavage of fluorophore moieties.

A time course analysis of the FRET signal, as a function of the time of incubation of cells with nocodazole showed that the loss of FRET signal parallels the MT disruption (Figures 3C,D), and that just 10 min after addition of nocodazole, MTs start to disassemble and NFRET dropped by almost 50%. This further demonstrates that CST can be used to monitor the state of association of Tau to cellular MTs.

We sought to determine whether loss of FRET signal is a consequence of MT disruption or only of the detachment of Tau from MTs, into a soluble pool. To this aim we exploited the MT stabilizing drug Paclitaxel (PTX), that while inducing a stabilization of MTs, competes with Tau for its binding to MTs (Kar et al., 2003; Makrides et al., 2004; Hernández et al., 2013). CST expressing cells were PTX treated for 10 min, after which cells were imaged for Tubulin immunofluorescence and CST fluorescence (Supplementary Figures S2B,C), showing that while the MT network is intact, Tau does no longer decorate MTs. Moreover, the CST is not processed in PTX treated cells (Supplementary Figure S2C). Under these conditions, no FRET signal can be observed (Figures 3E,F).

Next, we investigated Tau mobility by Fluorescence Recovery After Photobleaching (FRAP). FRAP technique allows to distinguish the freely diffusing pool of proteins, compared to the pool bound to other structures (Azoury et al., 2008; Himmel et al., 2009). The rate of fluorescence recovery indicates how fast neighboring fluorescent molecules fill the bleached area and the protein mobility is determined by both diffusion and binding interactions (Breuzard et al., 2013). We performed FRAP experiments in CST reporter cells, either untreated, or treated with Nocodazole or PTX. The corresponding recovery curves were fitted with two-exponential functions (Figure 3G, Supplementary Table S1 and Supplementary Figure S2D) and show that the CST FRAP curves in the Nocodazole and the PTX treated cells are identical, and different from the corresponding curve from untreated cells (Figure 3G).

We observed that 76% ± 3 of the CST molecules in untreated cells are in the mobile fraction contributing to fill the bleached area (Supplementary Table S1). The recovery curve analysis revealed that two main components are represented in this mobile fraction: the 27.8% ± 4.1, ascribed to diffusing soluble molecules (A₁) and the 72% ± 5, ascribed to a MTs-bound phase (A₂). The time constant associated with the rapid diffusion component is τ₁ of 4.9 ± 0.6 s while the fluorescence recovery due to MT-bound CST molecules occurred with a time constant which was seven-fold longer (τ₂ = 35.2 ± 2.8 s. These values are in agreement with previous studies, considering the presence of two fluorophores in each CST molecule (Breuzard et al., 2013).

The corresponding values for the Nocodazole and the PTX-treated cells show a large increase in the overall mobile fraction (91 ± 4 and 93 ± 3 respectively), with a two-fold increase in the diffusive fraction (53 ± 5 and 55 ± 3 respectively) and the concomitant reduction of the MT-bound fraction (47 ± 4 and 45 ± 3 respectively). The fact that we still observe two components with Nocodazole, while Breuzard et al. (2013) fit their FRAP Nocodazole curves with only one exponential might be due to incomplete disruption of the MT network, under our experimental conditions. These results provide a further demonstration that the Tau conformation in the soluble pool does not depend on the integrity of the MT network, but is rather an intrinsic property of soluble Tau molecules.

In conclusion, these results demonstrate that, in the cell, Tau molecule bound to MTs assumes a structured hairpin folding conformation, while it is an unstructured protein when not bound to MTs.

Tau point mutation at residue 301 is involved in FTDP-17 and is known to promote self-aggregation, but its contribution to the conformation of full length protein and to its mobility has not been clarified yet. Reporter cells expressing CST bearing P301L mutation have been analyzed by image analysis, FRET and FRAP (Figure 4). Fluorescence confocal imaging of CST-P301L reporter cells revealed a remarkable qualitative difference in MTs decoration by this Tau mutant, compared to CST, highlighted by magnification boxes (Figure 4A).

Image processing and analysis quantified the total length of Tau labeled segments on MTs and the number of crossover points of the fluorescent network (Figure 4B and Supplementary Figure S3D). Cells with comparable fluorescence intensity were analyzed. The results show, for both parameters, a significant decrease in the cells expressing the CST-P301L. This analysis indicates, for Leu mutant, an impaired ability of interacting with MTs. Conversely, the expression of these mutants does not affect the overall architecture of the MT lattice, nor the integrity and stability of the CST (Supplementary Figure S3A).

FRET analysis clearly demonstrated that P301 mutation altered Tau conformation on MTs, reducing by 50% the NFRET value (from 14.5 ± 0.5 to 8 ± 0.3; Figures 4C,D). This result indicates that P301L Tau mutant, even when it is associated to MTs, has an extended conformation that disadvantage FRET events.

Both image analysis and FRET experiments indicated that the P301L mutation strongly interferes with protein conformation and MTs interplay, therefore we investigated Tau mutant mobility by FRAP (Figure 4E).

FRAP curves revealed that mutant Tau proteins displayed an overall higher mobile fraction with respect to CST (85 ± 3% for P301L mutant), indicating that the P301 mutation increases Tau mobility (Figure 4E, Supplementary Figure S3, Supplementary Table S1). Remarkably, the rapid diffusion phase doubles for the CST-P301L (47 ± 4%) highlighting that the Pro to Leu mutation impacts on the balance between diffusive and binding phases. The time constants of both diffusing and bound phases are not altered, indicating that Tau molecule is not proteolytically processed, as confirmed by the western blot analysis (Supplementary Figure S3B).

We have further employed active and defective Tau mutants to investigate the conformational properties of Tau in relation to its interactions with MT. It is well known that phosphorylation of Tau affects its interaction with MTs, moreover, a fraction of CST is phosphorylated in HeLa cells (Supplementary Figure S3C, Plouffe et al., 2012). In order to prevent the phosphorylation at the AT8 epitope, that is known to occur at an early stage of the disease and to favor the phosphorylation of other pathological epitopes that are related to the MT destabilization, we generated the S199A-S202A-T205A AT8-epitope phosphorylation-defective mutant (CST-AT8mut; Supplementary Figure S3C, Braak and Braak, 1994; Braak et al., 1994; Bertrand et al., 2010; Bhaskar et al., 2010; Bibow et al., 2011b; Lippens et al., 2016). Conversely, we generated the defective Tau CST mutant ∆K280, also found in FTDP-17 patients and known to alter Tau interaction with MTs (Momeni et al., 2009; Hutton, 2000). Neither mutants showed alteration of the MT network (Supplementary Figure S3A).

The CST-AT8mut shows a FRET signal significantly higher than that of CST (Figures 4C,D), while the CST ∆K280 shows a FRET signal lower than that of CST and comparable to that of the P301L mutant.

FRAP measurements show that the FRAP recovery curves of the CST ∆K280 is very similar to that of the P301L mutant (Figure 4F and Supplementary Table S1), while that of the CST-AT8mut is significantly slower than that of CST, indicating a lower mobile fraction (54% vs. 76% overall mobile fraction, Supplementary Table S1).

Overall these results demonstrate that the FTDP-17 mutations P301L and ∆K280 consistently alter both the protein conformation and the interaction of Tau with MTs, conferring it an increased mobility. On the other hand, the phosphorylation-defective mutant AT8mut shows a greater binding to MTs, and, consistently, a higher FRET signal and a slower mobility.