. Tau PTM Profiles Identify Patient Heterogeneity and Stages of Alzheimer's Disease. Cell. 2020 Dec 10;183(6):1699-1713.e13. Epub 2020 Nov 13 PubMed.


Please login to recommend the paper.


  1. This is an impressive proteomic study providing a large list of tau post PTMs identified in different brain extracts including AD brain aggregates containing insoluble tau. The authors performed a semi-quantitative analysis measuring the frequency of detection of the different PTMs identified in brain extracts and the absolute quantitation of numerous tau peptides across the protein. To my knowledge, this is the largest brain cohort investigating AD tau by proteomics to date.

    It confirms previous reports on phosphorylation enrichment in AD insoluble tau mid-domain (i.e. pT181, pT217) and C-terminus (pS404). It also confirms extensive modification of the AD insoluble tau microtubule binding domain by methylation, acetylation, and ubiquitination.

    The FlexiTau method provides an indirect quantitation of tau modification across the protein. The PTMs’ abundance is inferred from the relative decrease in unmodified peptide levels compared to the other parts of the protein. There is some limitation to determining with precision the modification occupancy for each tau residue, and which PTMs contribute the most to the decreased level. Notably, the extent of modification reported could be affected in part by truncation.

    For example, an enrichment of the MTBR region (core of tau aggregation) and a partial degradation of the fuzzy coat (the soluble part surrounding the core) could contribute to a relative decrease in the abundance of the peptides still accessible to proteolysis in AD insoluble tau (Fitzpatrick, 2017). I note low molecular-weight tau characterized in this study is remarkably similar to the profile we measured in cerebrospinal fluid (CSF) using an analogous mass spectrometry (MS) quantitative approach (Barthélemy et al., 2016). As in CSF, N-terminus and mid domain are enriched compared to MTBR and C-terminus, likely due to degradation of the protein domain after 224 cleavage (Cicognola et al., 2019).

    An alternative MS method to measure tau modification occupancy is to quantify simultaneously modified and unmodified peptides. However, this approach is difficult to apply to all of the tremendous number of PTMs reported by the Steen group without considerable effort in designing MS assays for each reported PTM.

    We have currently applied this method to tau phosphorylation monitoring in brain, CSF, and plasma. We confirmed, for example, a significant enrichment of tau species phosphorylated on T181, S202, T217 or T231 in insoluble tau, consistent with the results reported by Wesseling et al.

    However, Wesseling et al did not find differences in soluble tau when AD and controls were compared. Conversely, we found hyperphosphorylation in soluble tau from AD brain on T111, T153, T205 or T217 (Horie et al., 2020). The same is observed in AD CSF (Barthélemy, 2019). These PTMs are typically modified in soluble brain tau and CSF with a low phosphorylation occupancy (<5 percent). Thus, discrepancy between the two reports could be due to differences in sensitivity between the two MS strategies employed.


    . Cryo-EM structures of tau filaments from Alzheimer's disease. Nature. 2017 Jul 13;547(7662):185-190. Epub 2017 Jul 5 PubMed.

    . Tau Protein Quantification in Human Cerebrospinal Fluid by Targeted Mass Spectrometry at High Sequence Coverage Provides Insights into Its Primary Structure Heterogeneity. J Proteome Res. 2016 Feb 5;15(2):667-76. Epub 2016 Jan 19 PubMed.

    . Novel tau fragments in cerebrospinal fluid: relation to tangle pathology and cognitive decline in Alzheimer's disease. Acta Neuropathol. 2019 Feb;137(2):279-296. Epub 2018 Dec 13 PubMed.

    . Regional correlation of biochemical measures of amyloid and tau phosphorylation in the brain. Acta Neuropathol Commun. 2020 Aug 27;8(1):149. PubMed.

    . Tau Phosphorylation Rates Measured by Mass Spectrometry Differ in the Intracellular Brain vs. Extracellular Cerebrospinal Fluid Compartments and Are Differentially Affected by Alzheimer's Disease. Front Aging Neurosci. 2019;11:121. Epub 2019 May 21 PubMed.

    View all comments by Nicolas Barthélemy
  2. This work represents the first unbiased and systematic effort to provide qualitative and quantitative profiling of the diversity of the Tau proteoforms, and how they change during the progression of AD. It illustrates the power of the bottom-up approach to large data analysis based on unbiased biochemical data acquisition to generate clinically meaningful clusters within the subject cohorts. It represents a significant advance that paves the way for future studies to decipher the Tau PTM code. It guides efforts to develop new working hypotheses, novel tools and assays to elucidate roles of PTMs in regulating Tau aggregation, seeding and pathology spreading. The authors should be commended for their efforts and courage to tackle complex, fundamental questions.

    This study reinforces the complexity of the Tau PTMs. It reaffirms our previous calls for embracing the complexity of Tau as a necessary step to deciphering the Tau PTM code and understanding its role in health and disease (Haj-Yahya and Lashuel, 2018; Haj-Yahya et al., 2020). This study yielded several interesting novel findings. Firstly, the cumulative number of the detected phosphorylation sites on Tau was much lower than the predicted number of sites, 55 out of possible 85. The clustering of the PTMs in specific domains, and the presence of multiple PTMs throughout the sequence, suggest that the Tau PTM code is a combinatorial code that involves complex crosstalk and the interplay between different PTMs in disease (Haj-Yahya and Lashuel, 2018; Haj-Yahya et al., 2020). In addition, it underscores the crucial point that Tau is not a single protein, but six unique proteins, in both their associations with differential pathology development, as well as their distinct PTM profiles.

    However, we emphasize that the PTM frequency and maps the investigators provided are still incomplete. Firstly, they were based on single-time-point snapshots in different patients, therefore the longitudinal alterations of PTM profiles are not known for any specific individual to robustly establish the temporal sequence of PTM occurrence. Secondly, maps represented the cumulative occurrence of the PTMs, which was derived from the detection of predominantly individual PTMs on the Tau peptide fragments. This lacks insight into the co-occurrence of PTMs along broader regions of the same protein molecule, and does not provide the information on the isoform origin, apart from several peptides containing isoform-differentiating regions at N- and C-terminus, but not for the domains common to all six isoforms. Furthermore, it is intriguing that several other PTMs reported to occur on Tau were not reported, discussed, or addressed in the proposed model. These include nitration, oxidation, tyrosine phosphorylation, O-GlcNacylation, as well as N-glycosylation. It is unclear whether these PTMs were absent in the samples, under the detection limit of the experimental methods, or the data analysis was not optimized for detection of these modifications. 

    It is difficult to gain insight into how the various PTM patterns evolve over time, or when the different PTMs are introduced during oligomerization and aggregation of Tau on the pathway to neurofibrillary tangle formation. The authors attempted to address this by correlating the PTM patterns and disease stages, but this remains a correlation based on a small number of cases.

    For example, the authors emphasized that the underrepresentation of modified N-terminal domain could suggest it may not significantly influence tau seeding activity, and thus antibodies targeting this domain are “predicted to be less-effective therapeutic agents for AD.” This interpretation illustrates the problem of reconstructing a process based on looking only at the end product. The existing data do not rule out important roles of PTMs in the N-terminal domain in the initiation and modulation of Tau oligomerization and aggregation, as well as its synergistic roles with other PTMs alongside Tau molecule (Ait-Bouziad et al., 2020).  It is also plausible that interactions between the N-terminus and central domains of Tau may influence their folding and the susceptibility to modification of the polyproline-rich region and MBD at the early stages of Tau aggregation.

    The authors also found that peptides bearing phosphorylated S262 were highly enriched in the sarcosyl-insoluble fractions of AD patients, which by the definition of the brain lysate preparation protocol used should not have contained NFTs or PHF-enriched Tau material (Mair et al., 2016; Greenberg and Davies, 1990). In vitro, we have shown that phosphorylation at S262 inhibited rather than promoted Tau aggregation process (Haj-Yahya et al., 2020), raising the possibility of enrichment of the pS262 Tau species in the sarcosyl-extracted fraction, not recruitment into NFT aggregates in AD.

    Two models were presented in this work. The first, proposing successive PTM addition with disease progression stages, is likely to apply only to the pool of relatively dynamic and semi-soluble sarcosyl-extractable Tau monomeric, oligomeric, or multimeric species. The study detected clustering of several PTMs in the microtubule-binding repeat domains. However, these modifications were not detected in the experimental data obtained from cryogenic electron microscopy coupled with mass spectrometry analyses of the structures of Tau filaments isolated from the brains of patients with AD and tauopathies (Arakhamia et al., 2020). This underscores the diversity in PTM patterns of different Tau species, at the level of Tau the monomer, oligomer, and high-order fibrillar assembly.

    Indeed, many of the PTMs detected in this work are expected to inhibit rather than promote Tau aggregation. For example, the authors explained the effects of acetylation and ubiquitination on the basis of charge neutralization. Introducing a bulky protein such as ubiquitin at a single or multiple lysine residue in the aggregation-prone MTBR domain is more likely to interfere with the onset and the process of Tau aggregation. This is what we have seen in our work with α-synuclein and mutant Huntingtin, where mono-ubiquitination was shown to stabilize the monomeric proteins and inhibit its aggregation.

    The second model suggests that phosphorylation or hyperphosphorylation of Tau drives its aggregation. This is not a universal phenomenon, as we and others have shown using site-specifically phosphorylated proteins. There are likely some phosphorylation patterns that promote aggregation, but this has to be further validated experimentally.

    Nonetheless, the data provided here represent an essential starting point for developing testable models, including the ones proposed by the authors. This could be achieved using the methodologies we recently developed for reconstructing Tau isoforms from scratch, and for producing homogenous proteins that are site-specifically modified at single or multiple residues. Using these methods, we were able for the first time to investigate the role of site-specific acetylation at K280 and investigate the role of cross-talk between different phosphorylation sites (Haj-Yahya and Lashuel, 2018; Haj-Yahya et al., 2020). The ability to generate these site-specifically modified proteins allows for more precise mapping of the role of PTMs in regulating Tau function in health and disease.

    The findings presented in this work must be independently validated in larger cohorts of patients, and across broader brain regions specifically associated with AD dementia, before we start to rethink our approaches to developing Tau-targeted diagnostics and therapeutics. Future studies are needed to refine the distribution maps of the Tau PTMs, with an emphasis on mapping co-occurring PTMs, as well as their Tau isoform origin, cross-talk between the PTMs in each domain, and the interplay between PTMs in different protein regions.

    This exciting work highlights that it is time we embrace the complexity of Tau. This is essential for bridging the widening gap between what is being studied in vitro using unmodified recombinant proteins, and the complexity of Tau species in the brain.


    . Phosphorylation of the overlooked tyrosine 310 regulates the structure, aggregation, and microtubule- and lipid-binding properties of Tau. J Biol Chem. 2020 Jun 5;295(23):7905-7922. Epub 2020 Apr 27 PubMed.

    . Posttranslational Modifications Mediate the Structural Diversity of Tauopathy Strains. Cell. 2020 Feb 20;180(4):633-644.e12. Epub 2020 Feb 6 PubMed.

    . A preparation of Alzheimer paired helical filaments that displays distinct tau proteins by polyacrylamide gel electrophoresis. Proc Natl Acad Sci U S A. 1990 Aug;87(15):5827-31. PubMed.

    . Site-Specific Hyperphosphorylation Inhibits, Rather than Promotes, Tau Fibrillization, Seeding Capacity, and Its Microtubule Binding. Angew Chem Int Ed Engl. 2020 Mar 2;59(10):4059-4067. Epub 2020 Jan 28 PubMed.

    . Protein Semisynthesis Provides Access to Tau Disease-Associated Post-translational Modifications (PTMs) and Paves the Way to Deciphering the Tau PTM Code in Health and Diseased States. J Am Chem Soc. 2018 May 30;140(21):6611-6621. Epub 2018 May 21 PubMed.

    . Single Posttranslational Modifications in the Central Repeat Domains of Tau4 Impact its Aggregation and Tubulin Binding. Angew Chem Int Ed Engl. 2019 Feb 4;58(6):1616-1620. Epub 2019 Jan 17 PubMed.

    . FLEXITau: Quantifying Post-translational Modifications of Tau Protein in Vitro and in Human Disease. Anal Chem. 2016 Apr 5;88(7):3704-14. Epub 2016 Mar 7 PubMed.

    View all comments by Galina Limorenko
  3. The extensive data in this article is an outstanding resource for the field. It nicely confirms several known features of the progression of tauopathy in AD and adds new information to refer to for ongoing and future studies.

    With regard to therapies, I agree, as we and others have previously discussed in articles on this topic, that a personalized treatment based on epitope profile that relates to the stage of the disease may be feasible in the future. However, I think the focus should be to target epitopes that are prominent in AD and other tauopathies, but are also found to some extent in normal subjects.

    As alluded to in this article, these are likely to be the earliest pathological epitopes and the control subjects who have them in low amounts may simply be in their early stages of developing tauopathy. Epitopes that are specific to AD/tauopathy are less attractive for targeting in large part because they occur late in the disease, as nicely documented in this article. These late epitopes are most prominent in insoluble tau, which is not very accessible to treatment and anyway probably better left alone.

    Lastly, the point the authors make about N-terminal tau not being a good target for therapy relates to what we have mentioned over the years, because this region has been shown previously by several laboratories to be missing in many forms of pathological tau. It is reasonable to target normal epitopes as long as they are also prominent in pathological tau.

    It will be interesting to see if the epitope profiles in this article will correlate with efficacy of the many tau immunotherapies in clinical trials. Given the complexity of the topic, the fact that antibodies against the same region can differ greatly in their efficacy, and considering that seeding and toxicity do not necessarily go hand in hand, the answer is unlikely to be straightforward.

    View all comments by Einar Sigurdsson
  4. This paper represents a major advance in the understanding of molecular changes in neuronal tau protein during the progression of Alzheimer‘s disease. This was made possible in large part by advances in mass spectroscopy methods, making use of calibration of peptide masses with regard to isotope-labeled peptides, embedded in the FLEXITAU procedures described by the authors previously (Mair et al., 2016; for recent updates on methods see Schlaffner et al., 2020). This was now applied to cohorts of AD patient brains and non-AD controls. The analysis revealed multiple modifications along the tau chain, mostly phosphorylation, acetylation, and ubiquitination, occurring in different tau isoforms, derived from different states of aggregation (tau monomers, oligomers, and fibers), and changing with the progression of the disease. As such, the paper presents a "high resolution" view into the Braak stages that were originally based on silver-stained neurofibrillary tangles (Braak and Braak, 1991). Wesseling et al. is in several ways complementary to recent studies by Barthélemy, Bateman and colleagues on tau protein from CSF (Barthélemy et al., 2019). 

    Here we will restrict our comments to a comparison with a recent study of tau protein by "native MS" to analyze intact, uncleaved tau, and combined with subsequent LC-MS/MS of tryptic peptides from the same protein (Drepper et al., 2020). Our study was done on the eukaryotic cell line Sf9, expressing the full-length human tau isoform (termed 2N4R, 441 residues, UniprotKB 10636-F). It revealed that tau, which was rapidly purified from cells to minimize enzymatic modifications, has a high extent of phosphorylation at Ser/Thr residues, ~8±5 Pi per tau molecule in a bell-shaped distribution, with states separated by 1 unit of Pi.

    Other modifications, i.e. phosphorylation at Tyr, acetylation, ubiquination, are minor, at or below the limit of detection. When phosphatases are inhibited by okadaic acid before cell preparation, the extent of phosphorylation rises to 14±6. Subsequent peptide analysis reveals ~51 phosphorylation sites out of 85 potential sites, with 90 percent sequence coverage. In agreement with Wesseling et al., the P-sites are preferentially located in the basic C-terminal half of tau, i.e. residues 150-441. Thus, much of the basic character of tau, caused by the dominant Lys+Arg residues, is neutralized by acidic phospho-Ser or -Thr. Phosphorylation is prominent at sites considered to be diagnostic for AD-tau, i.e. the epitopes of antibodies AT-8, AT-270, PHF-1 etc. Our conclusion is that phosphorylation is high and is the dominant modification of tau in a healthy eukaryotic cell. This reflects a strong phosphorylation "tone," as revealed by a natively unfolded protein like tau, in agreement with previous studies (Tepper et al., 2014; Mair et al., 2016). 

    How does this compare with the common notions in the AD field that tau is "hyperphosphorylated" in AD, and that hyperphorylation leads to fibrous aggregation of tau, i.e. paired helical filaments (PHFs), coalescing into neurofibrillary tangles (NFTs)? This view stems from early investigations showing that tau obtained from AD patient brains is highly phosphorylated (~10 Pi/molecule), in contrast to normal adult human brain tau (~2 Pi/molecule) (Ksiezak-Reding et al., 1992; Köpke et al., 1993; Kenessey and Yen, 1993). This led to the idea that hyperphosphorylation of tau is the cause of the pathologic conversion of soluble into aggregated tau at advanced age.

    As a consequence, many labs investigated phosphorylation sites, corresponding protein kinases, and protein phosphatases. The aim was to find early diagnostic markers of AD, and to develop therapeutic approaches to reduce tau aggregation, either by inhibiting the responsible kinases or by activating phosphatases. Early diagnostic markers were indeed found, most recently phospho-peptides of tau in CSF and blood (Bateman et al., 2020; Moscoso et al., 2020). However, modulation of tau phosphorylation via kinases or PPases has not led to treatments.

    It turns out that there may be a problem in the concept that pathological hyperphosphorylation leads to aggregation. Tau is an unusually hydrophilic and disordered protein that remains soluble in vitro at <200 µM irrespective of phosphorylation (Tepper et al., 2014). However, tau readily forms aggregates in the presence of polyanions like sulphated glycosaminoglycans or RNA of various origins (Goedert et al., 1996; Kampers et al., 1996). This suggests that polyanionic co-factors—rather than phosphorylation—could be the cause of aggregation. In fact, there has been a string of papers indicating that normal cellular tau is phosphorylated at many sites that were initially considered to be AD-specific (Matsuo et al., 1994; Sergeant et al., 1995; Gartner et al., 1998Wang et al., 2016). 

    Alas, this did not change the predominant view of pathological hyperphosphorylation causing tau aggregation. The likely explanation for the ostensibly low phosphorylation of normal adult tau lies in the fact that tau's state of phosphorylation is extremely sensitive to preparation conditions. As shown 12 years ago (Planel et al., 2007), even a small decrease in temperature during anesthesia generates enhanced phosphorylation because PPases lose activity faster than kinases. More importantly to AD, though, at time of death, kinases lose their activity more quickly because the co-factor ATP is depleted, whereas PPases continue to operate because they do not require ATP. As a result, tau from normal adult brains, prepared after a postmortem delay of many hours, is mostly dephosphorylated. Tau from fetal brain or from brain operations is prepared rapidly and is still in a high state of phosphorylation (Kenessey and Yen, 1993). By contrast, tau from AD brain remains in a high state of phosphorylation, most likely because many sites become inaccessible to PPases in tau‘s fibrous or oligomeric states (Schneider et al., 1999). 

    What conclusions can one draw from these observations?

    (1) There is room for doubt that tau phosphorylation is causally involved in the progression of AD, as most evidence is only correlative. This might explain the failure of treatments aimed at modulating kinases or PPases.

    (2) Certain phosphorylation sites of tau are valid early markers of AD, but not necessarily in a mechanistic sense; they might merely reflect the neuronal phosphorylation tone, and the protection of oligomeric or polymeric tau against PPases during the postmortem period. Thus, these early markers may help to identify other non-tau factors involved in the onset of AD, e.g. changes in energy metabolism, microglia activation, etc.

    (3) If the phosphorylation tone is altered in the course of AD, this will affect not only tau but multiple other proteins whose roles may have escaped attention so far. 

    (4) Phosphorylation is the major post-translational modification of tau in healthy cells, e.g. Sf9 cells known for their resistance to toxicity and radiation. As an intrinsically disordered protein with many Ser/Thr sites, tau is a faithful sensor of the phosphorylation tone, but the functional significance of sites is a different matter. Thus the involvement of multiple other modifications besides phosphorylation, occurring in different tau isoforms presented by Wesseling et al., 2020, and their integration into the cascade of AD progression, opens a new perspective for the identification of other, non-tau factors involved in the onset of tau aggregation and AD progression.

    Acknowledgements: We thank Dr. Emmanuel Planel (Laval Univ., Quebec) for insightful discussions.


    . Tau Phosphorylation Rates Measured by Mass Spectrometry Differ in the Intracellular Brain vs. Extracellular Cerebrospinal Fluid Compartments and Are Differentially Affected by Alzheimer's Disease. Front Aging Neurosci. 2019;11:121. Epub 2019 May 21 PubMed.

    . Another step forward in blood-based diagnostics for Alzheimer's disease. Nat Med. 2020 Mar;26(3):314-316. PubMed.

    . Neuropathological stageing of Alzheimer-related changes. Acta Neuropathol. 1991;82(4):239-59. PubMed.

    . A combinatorial native MS and LC-MS/MS approach reveals high intrinsic phosphorylation of human Tau but minimal levels of other key modifications. J Biol Chem. 2020 Dec 25;295(52):18213-18225. Epub 2020 Oct 26 PubMed.

    . Postmortem changes in the phosphorylation state of tau-protein in the rat brain. Neurobiol Aging. 1998 Nov-Dec;19(6):535-43. PubMed.

    . Assembly of microtubule-associated protein tau into Alzheimer-like filaments induced by sulphated glycosaminoglycans. Nature. 1996 Oct 10;383(6600):550-3. PubMed.

    . RNA stimulates aggregation of microtubule-associated protein tau into Alzheimer-like paired helical filaments. FEBS Lett. 1996 Dec 16;399(3):344-9. PubMed.

    . The extent of phosphorylation of fetal tau is comparable to that of PHF-tau from Alzheimer paired helical filaments. Brain Res. 1993 Nov 26;629(1):40-6. PubMed.

    . Microtubule-associated protein tau. Abnormal phosphorylation of a non-paired helical filament pool in Alzheimer disease. J Biol Chem. 1993 Nov 15;268(32):24374-84. PubMed.

    . Phosphate analysis and dephosphorylation of modified tau associated with paired helical filaments. Brain Res. 1992 Dec 4;597(2):209-19. PubMed.

    . FLEXITau: Quantifying Post-translational Modifications of Tau Protein in Vitro and in Human Disease. Anal Chem. 2016 Apr 5;88(7):3704-14. Epub 2016 Mar 7 PubMed.

    . Biopsy-derived adult human brain tau is phosphorylated at many of the same sites as Alzheimer's disease paired helical filament tau. Neuron. 1994 Oct;13(4):989-1002. PubMed.

    . Time course of phosphorylated-tau181 in blood across the Alzheimer's disease spectrum. Brain. 2021 Feb 12;144(1):325-339. PubMed. Correction.

    . Anesthesia leads to tau hyperphosphorylation through inhibition of phosphatase activity by hypothermia. J Neurosci. 2007 Mar 21;27(12):3090-7. PubMed.

    . FLEXIQuant-LF to quantify protein modification extent in label-free proteomics data. Elife. 2020 Dec 7;9 PubMed.

    . Phosphorylation that detaches tau protein from microtubules (Ser262, Ser214) also protects it against aggregation into Alzheimer paired helical filaments. Biochemistry. 1999 Mar 23;38(12):3549-58. PubMed.

    . Isoelectric point differentiates PHF-tau from biopsy-derived human brain tau proteins. Neuroreport. 1995 Nov 13;6(16):2217-20. PubMed.

    . Oligomer formation of tau protein hyperphosphorylated in cells. J Biol Chem. 2014 Dec 5;289(49):34389-407. Epub 2014 Oct 22 PubMed.

    . Corrigendum: Rapid alteration of protein phosphorylation during postmortem: implication in the study of protein phosphorylation. Sci Rep. 2016 Jan 7;6:17241. PubMed.

    View all comments by Eckhard Mandelkow

Make a Comment

To make a comment you must login or register.