Introduction

Tau is a microtubule-associated protein (MAP) mainly found in neurons and expressed inthe adult human brain as 6 isoforms (ranging from 352 to 441 amino acid residues inlength), which are derived from a single gene, MAPT, by the alternative splicingof exons 2, 3 and 101. Tau is composed of an amino terminal acidicdomain followed by two proline-rich domains and a microtubule-binding domain2. The latter contains 3 or 4 microtubule-binding repeats, dependingon whether the sequence encoded by exon 10 is included or not3. Tauis primarily involved in the regulation of microtubule stability and dynamics as well asaxonal transport4,5. Besides its role as a MAP, Tau exhibits othercellular localizations and functions that have been less investigated6,7,8. Tau proteins aggregate into filaments in a large group ofneurodegenerative disorders referred to as Tauopathies, such as Alzheimer'sDisease (AD) and Frontotemporal Dementia with Parkinsonism linked to chromosome 17(FTDP-17)9. AD is the most common Tauopathy and form ofdementia. One of its neuropathological hallmarks is neurofibrillary degeneration (NFD),characterized by aggregated Tau proteins. Studies have shown that the progression of NFDin cortical brain areas is closely correlated to cognitive impairment in AD, supportinga central role for Tau in AD pathology10,11. As of now, themechanisms leading to NFD and its progression are far from being elucidated.Nevertheless, the deregulation of Tau phosphorylation is a key event in the pathologicalprocess. Numerous studies suggest that abnormal phosphorylation impedes Tau binding tomicrotubules, leading on the one hand to the depolymerization and loss of the latter,and on the other hand to the formation of toxic aggregated Tau species12. Truncation is another post-translational modification that could have anetiological role in Tau pathology. Numerous cell-based assays show that the truncationof either the C-terminal part of Tau or both the N- and C-terminal parts impacts itsbiochemical and functional properties and triggers a gain of toxic function13,14,15,16. Moreover, animal models based on the expression ofparticular truncated Tau species are able to reproduce Tau pathology16,17,18. The analysis of AD brains by western blotting (WB) andepitope mapping suggests the occurrence of cleavage sites in both the N-terminal andC-terminal parts of Tau proteins19,20. While several N- andC-terminally truncated Tau species are observed in AD brains, only a limited number ofspecific Tau cleavage sites, after residues Asp13, Asp25, Asn368, Glu391 and Asp421,have been identified so far in situ. The species generated by these cleavages arefound in neurofibrillary tangles and their occurrence is correlated with the severityof the disease16,21,22,23,24,25,26. The preciseidentification of new Tau cleavage sites is a mandatory step in the generation ofappropriate experimental tools with which to investigate their impact on Taufunction/dysfunction and obtain new insights into the role of Tau truncation inpathological process but also in physiological framework. In this context, we undertooka challenging proteomics approach to precisely identify new Tau cleavage sites,especially those at the N-terminus, which are less well characterized than those at theC-terminus. We have identified several new N-terminally truncated Tau species in thehuman brain, with N-terminal residues located throughout the Tau sequence, leading us toexpect that these Tau species would be of crucial functional and/or pathologicalrelevance. We therefore initiated cell-based functional studies by analyzing thebiochemical characteristics of two N-terminally truncated Tau species. Our resultssuggest, surprisingly, that the Tau N-terminus could play a direct role in theregulation of microtubule dynamics.

Results

Identification of new N-terminally truncated Tau species in the humanbrain

WB analyses of human cortical brain samples using antibodies directed against theN- and C-terminal parts of Tau protein revealed several truncated Tau species(Fig. 1A), in agreement with previous reports19,20. These Tau species are also showed by using pSer396antibody (Fig. 1B). It is worth noting that, as expected,differential phosphorylation of the various Tau species was observed between thesamples and between the different brain areas with respect to Braak stages (7).In order to pinpoint new N-terminal truncation sites of Tau, we optimized aproteomic approach (Fig. 1C) in brain samples. The wholerange of Tau species was immunoprecipitated (IP) from protein extracts with theTau-5 antibody, which recognizes amino acid residues 218–22527. As shown by WB analysis (Fig. S1),Tau-5 IP allows the purification of full-length Tau (FL-Tau) as well as N- andC-terminally truncated Tau and aggregated species, which display lower andhigher molecular weights, respectively. In order to improve detectionsensitivity, Tau-5 IP products were then subjected to primary amine labelingusing a covalently-linked biotin prior to enzymatic digestion with eithertrypsin or the endoproteinase Asp-N. Labeled peptides were then purified onstreptavidin columns and identified using liquid chromatography-massspectrometry (LC-MS/MS). This approach resulted in the identification of anumber of Tau peptides (Table S1; data compiled from all samples), includingsome semi-tryptic/semi-Asp-N peptides. Indeed, 21 peptides displayednon-tryptic/non-Asp-N residues at the N-terminus and constituted our set ofcandidates for N-truncated Tau species (Table 1). Out ofthem, only 2 (Lys310, Lys 395) can be explained by the activity of possiblechymotrypsin contaminant. It should be noted that not all the peptides wereN-terminally labeled, in part because of the post-translational modification ofamines. Among these N-terminal sites, Ala2 was observed to beNα-acetylated, as previously reported by Hasegawa et al., who foundthis Tau species in both normal and AD brains28.

Table 1 Identification of 21 N-terminal truncation sites of Tau protein by LC-MS/MS.Semi-tryptic and semi-Asp-N peptides detected in several samples are shown, withthe corresponding first amino acid residue and N-terminal modifications(numbering of N-terminal residues correspond to the N-terminal cleavage siteidentified). In bold, identified cleavage sites on the N-terminal side of theTau-5 epitope
Full size table
Figure 1
figure 1

Identification of 21 N-terminal truncation sites of Tau protein from humanbrain tissue using LC-MS/MS.

(A): Characterization of human brain tissue by WB using antibodies directedagainst the Tau C-terminal (Tau-Cter) and N-terminal (Tau-Nter) ends;representative analysis of tissue from the frontal (F) and occipital (O)cortex of Braak 0 (B0), Braak III (BIII) and Braak VI (BVI) patients. GAPDHwas used as a loading control. (B): Characterization of the same human braintissue by WB using pSer396 antibody. The gels displayed in A and B have beenrun under the same experimental conditions. Cropped blots are displayed;Full-length blots are presented in supplementary data (as Fig.S7A and Fig. S7B respectively). (C):Proteomics approach developed to identify N-terminal sites of Tau protein;Tau species were immuno-enriched from the human occipital and frontalcortex, labeled with covalently-linked biotin, digested either with trypsinor with Asp-N and analyzed by LC-MS/MS. (D): Representation of the positionof identified cleavage sites as well as of the Tau-5 antibody epitope on aschematic Tau sequence (numbering according to the longest Tau isoform).

Full size image

These new N-terminal sites were scattered across the Tau sequence (Fig. 1D) and except for Ala2, have not been described before. TheN-terminal sites located C-terminally to the Tau-5 epitope were detectedprobably because they are found in situ in protein complexes with other Tauspecies, as suggested by WB analysis showing high-molecular-weight Tau species(Fig. S1). It is worth noting that known Tau bindingpartners such as Hsc7029 were also found in our analysis,indicating that the experimental conditions supported co-IP (Fig.S2).

The Gln124-Tau fragment shows a different phosphorylation pattern fromFL-Tau

Regarding the functional relevance of the new identified sites, since they arescattered widely across the Tau sequence, one could expect each cleavage site tohave a specific impact, whether exclusive or not, on Tau function andbiochemical properties. We therefore initiated functional studies by analyzingthe biochemical characteristics of Tau species starting at residues Met11 (Fig. S3A) and Gln124 (Fig. S3B). Theseresidues are located in the Tau projection domain and are present in all Tauisoforms (Fig. 1D). We generated expression vectorscontaining coding sequences (cDNAs) of FL- and truncated-Tau species (Fig. 2A). These constructs were transfected into the N1E-115neuroblastoma cell line. Forty-eight hours after transfection, WB analysis usingan antibody directed against the C-terminal part of Tau showed immunoreactivebands at the expected molecular weights as well as mobility shifts likelyrelated to particular phosphorylated Tau species (Fig.2B), but we cannot exclude that the shifts are due to othermodifications. An evaluation of phosphorylation status using antibodies tophospho-Tau showed differential phosphorylation patterns (Fig. 2Cand 2D). Compared to FL-Tau, Met11-Tau displayed an increase inphosphorylation at the Thr231 epitope (detected by the AT180 antibody), althoughno difference in phosphorylation was observed at Ser396. Interestingly,Gln124-Tau displayed a decrease in phosphorylation at Thr231 and Ser262/356(detected by the 12E8 antibody), compared to FL-Tau.

Figure 2
figure 2

Expression and phosphorylation pattern of truncated Tau proteins.

(A): Schematic representation of 1N4R FL-Tau isoform, which includes exons 2and 10 and the Met11-Tau and Gln124-Tau fragments. PR: proline rich domain.(B): Representative WB analysis using the Tau-Cter antibody of proteinextracts from N1E-115 cells transfected with control vector (mock), FL-Tauand the Met11-Tau and Gln124-Tau fragments. GAPDH was used as a loadingcontrol. (C-D) Representative WB analysis and densitometric quantificationsof phosphorylated epitopes (AT180: pThr231; 12E8: pSer252-pSer356 andpSer396). Quantification was performed by calculating the ratio ofphosphorylated Tau to total Tau (Tau-Cter), both relative to GADPH. Errorbars indicate SEM. N ≥ 3 independent experiments. *: P≤ 0.05; **: P ≤ 0.01. Differences between mean valueswere determined using One-way ANOVA followed by Fisher's LSD posthoc test. The gels displayed in B and C has been run under the sameexperimental conditions. Cropped blots are displayed; Full-length blots arepresented in supplementary data (as Fig. S8A and S8B, respectively).

Full size image

Cells expressing Gln124-Tau display an increase in α-tubulinacetylation and detyrosination

Since it has been shown that phosphorylation at Ser262/Ser356 strongly decreasesthe binding of Tau to microtubules30,31, we asked whetherthe truncation of the N-terminal part of Tau could have any consequences for Taufunction in microtubule stabilization. To investigate the properties ofN-terminally truncated Tau species with regard to microtubules, we firstcompared the effect of their expression on microtubule modifications, inparticular the acetylation of α-tubulin, which is a hallmark ofstable microtubules32. WB analysis of transientlytransfected N1E-115 cells showed that, as expected, cells expressing FL-Taudisplayed an increase in the level of α-tubulin acetylation (Fig. 3). Regarding the effect of N-terminally truncated Tau,the Met11-Tau fragment showed the same properties as FL-Tau while Gln124-Tau wasassociated with a markedly increased level of acetylated tubulin compared toFL-Tau. Given that Tau can interact with and inhibit the function of HDAC6, themain tubulin deacetylase33, we investigated whether theincrease in tubulin acetylation in cells expressing Gln124-Tau was related to astronger inhibition of HDAC6 activity. We measured HDAC activity in cytoplasmicextracts, since HDAC6 is mainly cytoplasmic compared to other HDAC members. Ourdata (Fig. S4) show that there was no significant differencein HDAC activity between cells transfected with FL-Tau and those withN-terminally truncated species, indicating that the effect on tubulinacetylation is not related to differences in HDAC6 activity. Instead, thiseffect could be a consequence of improved microtubule polymerization and/orstabilization. In agreement with this possibility, the investigation of anotherpost-translational modification of tubulin, detyrosination, which is alsorepresentative of a stable population of microtubules34,revealed that it was significantly increased with Gln124-Tau when compared toFL-Tau (Fig. 3).

Figure 3
figure 3

Gln124-Tau increases α-tubulin acetylation anddetyrosination.

(A): Post-translational modifications of α-tubulin analyzed by WBusing protein extracts from N1E-115 cells overexpressing FL-Tau, Met11-Tauor Gln124-Tau. The gels have been run under the same experimentalconditions. Cropped blots are displayed; Full-length blots are presented insupplementary data (as Fig. S9). (B): Quantification wasperformed by calculating the ratio of modified tubulin to total tubulin,both relative to GAPDH. Error bars indicate SEM. N ≥ 5independent experiments. *: P ≤ 0.05; **: P ≤ 0.01.Differences between mean values were determined using One-way ANOVA followedby Fisher's LSD post hoc test.

Full size image

Gln124-Tau binds more strongly to microtubules than FL-Tau and protectsagainst depolymerization

Since tubulin acetylation and detyrosination occur once microtubules arepolymerized, the increase in their levels in cells expressing Gln124-Tau islikely a consequence of improved microtubule polymerization and/orstabilization. To investigate the capacity of Tau species to promote microtubulepolymerization in cells, we analyzed their ability to induce neurite-likeextensions in response to cytochalasin B. The process formation in thisexperimental test is an indicator of the ability of Tau to assemble andpolymerize microtubules. Forty-eight hours after transfection, N1E-115 cellswere subjected to cytochalasin B treatment for 1 h. In agreement withwell-established reports35, immunolabeling analysis byconfocal microscopy showed that the breakdown of the cortical actin network bycytochalasin B allowed cells expressing FL-Tau to lengthen microtubule bundlesinto cellular extensions that looked like neurites (Fig.S5A). An analysis of cells expressing N-terminally truncated speciesshowed no evident difference in the number of cells displaying cellularextensions compared to cells expressing FL-Tau (Fig. S5B).It is worth noting that, in accordance with the reported in vitrostudies, the ability of Tau to assemble and polymerize microtubules requires thedomains involved in microtubule binding and assembly: the second proline-richdomain, the microtubule-binding repeats as well as inter-repeat regions36,37. Indeed, in our experiments, cells overexpressing aTau fragment that is truncated in the second proline-rich domain and a Taufragment that is truncated in the first repeat domain do not display any processformation (Fig. S6).

Next, we examined whether there was any difference in microtubule stabilization.We first compared the microtubule-association properties of FL-, Met11- andGln124-Tau species. Transiently transfected cells were fractionated into acytosol fraction (containing free Tau) and a microtubule fraction (containingmicrotubule-associated Tau) and were analyzed by WB (Fig.4A). As expected, both acetylated and detyrosinated tubulin formswere mainly present in the microtubule-enriched fraction. Immunolabeling with aTau antibody revealed that all Tau species were found in both microtubule andcytosol fractions. Nevertheless, the proportion of Gln124-Tau was higher in themicrotubule-enriched fraction (Fig. 4B), suggesting thatthis fragment binds more strongly to microtubules compared to FL-Tau orMet11-Tau.

Figure 4
figure 4

Gln124-Tau binds more efficiently to microtubules than FL-Tau.

(A): Representative WB analysis of microtubule fractions from N1E-115 cellextracts transiently transfected with FL-Tau, Met11-Tau or Gln124-Taufragments. The purity of the fractions was evaluated using an antibody toacetylated α-tubulin. The gels have been run under the sameexperimental conditions. Cropped blots are displayed; Full-length blots arepresented in supplementary data (as Fig. S10). (B):Quantification was performed by calculating the ratio ofmicrotubule-associated Tau to total Tau. Error bars indicate SEM. N≥ 3 independent experiments. *: P ≤ 0.05. Differencesbetween mean values were determined using One-way ANOVA followed byFisher's LSD post hoc test.

Full size image

We then tested the ability of Gln124-Tau to protect microtubules fromdepolymerization induced by nocodazole treatment. The nocodazole-resistanceexperiment allows the capacity of Tau to protect microtubules fromdepolymerization once they are assembled to be evaluated and is hence anindicator of microtubule stability and dynamics. Transfected N1E-115 cells weretreated for 20 minutes with nocodazole and subjected to cytosol/microtubulefractionation for WB analysis of microtubule depolymerization and Tau releasefrom microtubules. Nocodazole treatment induced microtubule depolymerization andTau release from microtubules, as shown by decreased tubulin and Tauimmunolabeling, respectively, in the microtubule fraction from treated cells(Fig. 5A). However, cells expressing Gln124-Tau weremarkedly more resistant to nocodazole treatment compared to cells expressingFL-Tau. Indeed, higher levels of tubulin were observed in the microtubulefraction from nocodazole-treated Gln124-Tau cells compared to cells transfectedwith the other Tau species (Fig. 5A and 5B). To confirmthese data, transfected N1E-115 cells were treated for 20 minutes withnocodazole and the disruption of Tau-induced microtubule bundles analyzed byconfocal microscopy (Fig. 5C). The results show thatunlike cells expressing FL-Tau or Met11-Tau, microtubule bundles were stillpresent in almost all cells expressing Gln124-Tau after 20 min oftreatment (Fig. 5C and 5D). Hence, microtubules in cellsexpressing Gln124-Tau are less sensitive to depolymerization than microtubulesin cells expressing FL-Tau or Met11-Tau.

Figure 5
figure 5

Gln124 protects cells more effectively against microtubule depolymerizationthan FL-Tau.

(A): Representative analysis of Tau species and total tubulin distribution inmicrotubule fractions, performed after 20 minutes of nocodazole treatment.(B): Quantification was performed by calculating the ratio of tubulinepresent in the microtubule fraction to total tubuline. The Cropped blots aredisplayed; Full-length blots are presented in supplementary data (as Fig. S11). (C): Confocal imaging of N15-115 cellstransfected with Tau and the truncated species Met11-Tau and Gln124-Tau, andtreated with nocodazole. Cells harbor bundles are designated by whitearrows. Scale bar: 50 μM. (D): Quantification of cells expressingTau, Met11-Tau or Gln124-Tau which display bundles after 20 minutes ofnocodazole treatment. Error bars indicate SEM. N ≥ 3 independentexperiments. ***: P ≤ 0.001. Differences between mean values weredetermined using One-way ANOVA followed by Fisher's LSD post hoc test.

Full size image

Discussion

In this work, using a dedicated proteomics approach, we identified new N-terminallytruncated Tau species in the human brain by identifying their precise primary amineresidues. Our finding lays the groundwork for the development of appropriate toolswith which to monitor species truncated at these newly identified sites underphysiological and pathological conditions and obtain new insights into the role oftruncation in Tau function and dysfunction.

Truncation is one of the post-translational modifications of Tau encountered in ADbrains, as shown by the detection of several Tau species truncated at the N- andC-terminus19,20. The determination of the precise cleavagesites of some of these Tau species was initially made possible by epitopemapping21,22,23,24,25. While the main C-terminalcleavage sites (Glu391 and Asp421) are well characterized, among the plethora ofN-terminally truncated species, only two have been identified in situ24,25. The purpose of our proteomics approach was to simplifypeptide mixtures before MS/MS identification, in order to enhance the probability ofidentifying the N-terminal extremities of these Tau species. Our approach requiredthe labeling of available amino groups on intact proteins using anN-hydroxysuccinide (NHS) ester derivative of biotin. Biotinylated peptides were thenpurified using streptavidin and analyzed by MS/MS. Indeed, as we started with humanbrain tissue, we needed to avoid multiple peptide-purification steps in order tostrike a balance between peptide loss and N-terminal enrichment, although effectiveprotocols have been published38. In order for a peptide to bevalidated as a candidate for N-terminal Tau truncation, its amino-terminal extremityshould not be a cleavage site for the protease used (trypsin or Asp-N) and it shouldbe found in several samples in order to reduce the probability of random cleavage bythe enzyme, even if trypsin has been reported to be very specific39. It should be noted that labeling of the amino-terminal group was notconsidered an absolute validation criterion due to the fact that our labelingprocedure was probably not complete under the experimental conditions used. For somepeptides, the amino-terminal group was blocked by post-translational modifications,such as the Nα-acetylation of Ala2. Here, we identified a number of newTau N-terminal sites, but there is a gap between the detection and thequantification of these Tau species in normal and AD brains. Right now, we cannotsay with any certainty which cleavage sites are relevant to the disease process andwhich are specific to the physiological context. Likewise, the impact of post-mortemdelay cannot be unequivocally addressed. The present approach allows for theidentification but not the quantification of the amino-truncated species. In fact,to obtain reliable quantitative results, it is necessary to avoid the IP andN-terminal labeling of Tau proteins we used in our study. To reach this goal, themethod of choice would be the development of specific antibodies, as was initiallydone for C-terminally truncated Tau species, especially those ending at Glu391 andAsp42121,22,23. Hence, our interesting findings providethe basis for the development of appropriate tools with which to monitor truncatedspecies that start at these newly identified sites, as well as the correspondingC-terminally truncated species, under physiological and pathological conditions.Moreover, these tools would be of interest in the diagnostic field.

Regarding the consequences of truncation to Tau protein properties, our initialcell-based studies show that the ability of Tau to stabilize microtubules is greaterwhen the N-terminal part is truncated. Indeed, the Gln124-Tau fragment showed astronger ability to bind microtubules and protect them from depolymerizationcompared to FL-Tau. In agreement with this supposition, cells expressing theGln124-Tau fragment display a significant increase in post-translationalmodifications characteristic of stable microtubules (α-tubulinacetylation and detyrosination). It is widely accepted, based mainly on invitro analyses of Tau fragments generated by amino-terminal deletions, thatTau binds microtubules and regulates their stabilization and polymerization throughits C-terminal part. These earlier studies indicated that the direct effects of Tauwith regard to microtubules involves a region encompassing amino acid residues215–358, which contains the second proline-rich domain, themicrotubule-binding repeats as well as inter-repeat regions36,37. The role of the Tau amino-terminal domain with regard to microtubules hasbeen reported as being indirect, such as by the regulation of microtubulespacing40 and functions41,42.Nevertheless, in lines with our data, in vitro studies of the impact ofmissense mutations encountered in Tauopathies (mutations at the Arg5 and at Gly55residues) suggest that the modification of the amino-terminal domain of Tau coulddirectly impact microtubules43,44,45. Besides, a recent invitro attempt to improve mechanisms of Tau interaction with microtubulesbased on the use of Tau fragments generated by limited proteolysis has shown thatthe Tau fragment Ser208-Ser324 binds more tightly to microtubules than FL-Tau andfavors their assembly46. In agreement with these invitro assays, our cell-based study of the N-terminally truncated Taufragment (Gln124-Tau) newly identified in situ suggests that theamino-terminal domain of Tau could directly regulate its binding and stabilizationof microtubules. To further characterize the Gln124-Tau fragment, it would be ofinterest to evaluate, on the one hand, whether the observed effects areisoform-dependent and on the other hand, the impact of Gln124-Tau on the functionsof FL-Tau. Indeed, the current work was performed in a cell line that does notdisplay detectable levels of endogenous FL-Tau.

Regarding the mechanisms underlying the gain of function displayed by the Gln124-Taufragment, one explanation could be related to the fact that the Tau protein is proneto adopt a “paperclip” conformation as a result ofintra-molecular interactions between the N-terminal and C-terminal domains47,48. Hence, N-terminal truncation would be expected to unfoldTau from this conformation and to expose the microtubule-binding domain. Thisexplanation is unlikely under our experimental conditions, since we find no obviousdifference with regard to microtubule stabilization between the Met11-Tau fragmentand FL-Tau. A more plausible explanation would be that Gln124-Tau, due to thetruncation of the negatively charged N-terminus, displays enhanced binding to thenegative surface of microtubules.

Concerning the biological significance of this gain of function, sustainedmicrotubule stabilization is likely to have a deleterious effect on neurons byimpairing synaptic plasticity and microtubule-dependent transport. Indeed, mutationsin FTDP-17 that lead to an increase in 4R Tau isoforms, which stabilize microtubulesmore strongly than 3R isoforms, are the cause of neuronal death and dementia49. Moreover, given that the microtubule-severing proteinskatanin and spastin have a more potent effect on stable microtubules50,51, a sustained increase in microtubule stability and itsassociated α-tubulin acetylation and detyrosination might lead to theincreased recruitment of severing enzymes and microtubule loss due to excesscutting52,53.

In summary, this work identifies new N-terminally truncated Tau species that occur inthe human brain and that have potential relevance to Tau biology and likely toAD-related Tauopathies. Our preliminary cell-based studies suggest, interestingly,that the N-terminal part of Tau could be directly involved in the regulation ofmicrotubule stabilization. N-terminal truncation may also influence other Tauproperties such as polymerization22 and cellularlocalization54. Future investigations based on our newlyidentified N-terminally truncated Tau species, in particular Gln124-Tau, shouldimprove our knowledge as to the role of truncation in Tau biology as well as in theAD pathological process.

Methods

Human tissue samples

Human brain autopsy samples were from the Lille NeuroBank collection (Centre deRessources Biologiques du CHRU de Lille). Informed consent was obtained from allsubjects. The Lille NeuroBank has been declared to the French Research Ministryby the Lille Regional Hospital (CHRU-Lille) on August 14, 2008 under thereference DC-2000-642 and fulfils the criteria defined by French Law regardingbiological resources, including informed consent, ethics review committeeapproval and data protection. The study was approved by the ethics reviewcommittee of Lille NeuroBank. The Table 2 indicates thestage of Tau pathology, categorized on the basis of neuropathologicalcharacteristics according to Braak and Braak11.

Table 2 Summary of brain tissues used for LC-MS/MS analysis
Full size table

Cell culture and transfection

N1E-115 mouse neuroblastoma cells were grown in Dulbecco's modifiedEagle's medium supplemented with 10% fetal calf serum withoutpyruvate, 2 mM L-glutamine and 50 units/mlpenicillin/streptomycin (Invitrogen) in a 5% CO2 humidified incubatorat 37°C.

Transfection with plasmid constructs was performed 24 hours after cell seedinginto six-well plates (for biochemical studies) or on culture slides (forimmunocytochemistry), using ExGen500 (Euromedex) according to themanufacturer's instructions, for 48 hours.

Plasmid constructs

Expression vectors carrying cDNA for FL and N-terminally truncated Tau 1N4Risoform were generated using the In-Fusion cloning Kit (Clontech) and PCRprimers were designed to clone inserts into the EcoRI site of pcDNA3.1(Invitrogen). Each cDNA fragment was amplified by PCR (DyNAzymeTM EXT DNApolymerase, New England BioLabs) from pcDNA3.1-Tau4R55.Forward primers were designed to contain the Kozak consensus sequence and are asfollows:

For FL-Tau,5′-CAGTGTGGTGGAATTCGCCACCATGGCTGAGCCCCGCCAGGAGTT-3′;

For Met11-Tau,5′-CAGTGTGGTGGAATTCGCCACCATGGAAGATCACGCTGGGACGT-3′;

For Gln124-Tau,5′-CAGTGTGGTGGAATTCGCCACCATGCAAGCTCGCATGGTCAGTAAAAGCAAAGACGGG-3′;

For Val229-Tau,5′-CAGTGTGGTGGAATTCGCCACCATGGTCCGTACTCCACCCAAGTCGCCGTCT-3′;

For Gly261-Tau,5′-CAGTGTGGTGGAATTCGCCACCATGGGCTCCACTGAGAACCTGAAGCACCAGC-3′.

The reverse primer for all amplifications was:

5′-GATATCTGCAGAATTCTCACAAACCCTGCTTGGCCAGGG AGGCA-3′.

DNA sequencing was carried out for construct validation.

Protein extractions

Tissues were homogenized by sonication in a buffer containing 0.32 Msucrose and 10 mM Tris-HCl, pH 7.4. For biochemical characterizationand immunoprecipitation, samples were prepared using IP-buffer:100 mM NaCl, 110 mM KOAc, 0.5% Triton X-100, BufferingSalts pH 7.4 (Invitrogen) with protease inhibitors (Complete w/o EDTA, Roche),sonicated and centrifuged at 2500 × g for 5 min.

For cell protein extracts; cells were washed using PBS and harvested in ice-coldRIPA buffer: 150 mM NaCl, 1% NP40, 0.5% sodium deoxycholate, 0.1%SDS, 50 mM Tris-HCl, pH 8.0, completed with protease inhibitors.Protein concentrations were determined using the BCA Assay Kit (Pierce).

Immunoprecipitation (IP)

IP was performed using the Dynabeads Co-Immunoprecipitation Kit 143.21D(Invitrogen). Briefly, the first step consisted of coupling 25 mg ofmagnetic beads with 250 μL of Tau-5 antibody (Invitrogen)according to the manufacturer's protocol. Then,1250 μg of protein extract were added to the bead-antibodycomplex and incubated overnight at 4°C. After washing with IP buffer,IP products were obtained by adding the elution buffer: buffering salts pH 2.8(Invitrogen). A small aliquot of this fraction was analyzed using WB and therest lyophilized using a speed-vac for proteomics analysis.

Identification of peptides by capillary liquid chromatography-tandem massspectrometry (LC-MS/MS)

IP Tau solubilized in PBS (pH 8.0) was reduced and alkylated (DTT,10 mM final, 2 h, 37°C; iodoacetamide,50 mM final, 30 min in the dark). After 3 washes with PBSon a microcolumn (Millipore, cut off 10 kDa), primary amines werelabeled with NHS-biotin (Thermo scientific) according to themanufacturer's instructions (1 mM, 2 h, RT).Then, NHS was saturated using 200 mM hydroxyalamine and proteinswashed 3 times with ammonium bicarbonate (50 mM). The protein extractwas then divided in two parts, one for proteolysis using trypsin(10 ng modified sequencing-grade trypsin; Roche; 37°C,overnight) and the other for proteolysis with Asp-N (10 ng, Roche;37°C, overnight). The resulting peptides were then purified usingstreptavidin beads (Thermo scientific) according to themanufacturer's recommendations. 300 μL ofresins were incubated for 1 h at RT with the samples. After washingwith PBS, bound peptides were eluted with DTT (5 mM final,30 min, RT), acidified with 10 μL of 10%aqueous formic acid and desalted using ZipTip C18 microcolumns (Millipore).

Peptides (2 μL) were purified on a capillary reversed-phasecolumn (nano C18 Acclaim PepMap100 Å, 75 μmi.d., 15 cm length; Dionex) at a constant flow rate of220 nL/min, with a gradient of 2% to 40% of buffer B in buffer A, for45 min; buffer A: water/acetonitrile/formic acid 98:2:0.1(vol/vol/vol); buffer B: water/ACN/formic acid 10:90:0.1 (vol/vol/vol). Couplingwith the mass spectrometer was done using a NanoMate (Advion; Voltage:+1.70 kV; Spray Sensing Enabled; Below 10.0 nA or above2000.0 nA for 5 secs; Chip ID: A8E159AK). The MS analysiswas performed on a Fourier transform ion cyclotron resonance (FT-ICR) massspectrometer (LTQ-FT Ultra; ThermoFisher Scientific) with the top-sevenacquisition method: MS resolution 60,000, mass range470–2,000 Da, followed by MS/MS (LTQ) of the seven mostintense peaks, with a dynamic exclusion of 90 s.

The raw data were processed using Xcalibur 2.0.7 software. The database searchwas done using the Mascot search engine (Matrix Science Mascot 2.2.04) on ahomemade protein databank containing the 6 human Tau isoforms and somecontaminants, or on the entire SwissProt protein databank. Proteome Discoverer1.3 (ThermoFisher Scientific) and Mascot were used to search data and filter theresults. The following parameters were used: MS tolerance 5 ppm;MS/MS tolerance 0.5 Da; semi-tryptic or semi-Asp-N peptides; onemissed cleavage allowed; partial modifications: carbamidomethylation (C),oxidation (M), phosphorylation (ST), acetylation (N-term), thiopropionation(N-ter, K).

The mass spectrometry proteomics data have been deposited to the ProteomeXchangeConsortium (http://proteomecentral.proteomexchange.org) via the PRIDE partnerrepository56 with the dataset identifier PXD001353 andDOI 10.6019/PXD001353.

Western blotting (WB)

Protein extraxts were standardized at 1 μg/μLwith LDS 2X supplemented with a reducing agent (Invitrogen) and denatured at100°C for 10 min. Proteins were then separated withSDS-PAGE using precast 4–12% Bis-Tris NuPage Novex gels (Invitrogen).Proteins were transferred to 0.45 μM nitrocellulosemembranes (AmershamTM Hybond ECL), which were saturated with 5%dry non-fat milk in TNT buffer; 140 mM NaCl, 0.5% Tween20,15 mM Tris, pH7.4, or 5% bovine serum albumin in TNT (Sigma)depending on the primary antibody. Membranes were then incubated with theprimary antibodies (Table S2) overnight at 4°C, washed with TNT threetimes for ten minutes, incubated with the secondary antibodies (Vector) andwashed again before development. Immunolabeling was visualized usingchemiluminescence kits (ECLTM, Amersham Bioscience) on anLAS-3000 acquisition system (Fujifilm). Labeling was quantified with ImageJsoftware (Scion Software).

Immunocytochemistry

Cells were plated on culture slides (BD Biosciences) coated with poly-D-lysine(Sigma). Twenty-four hours after plating, cells were transfected according tothe protocol previously described. Two days post-transfection, cell weresubjected to the appropriate treatment: 20 μM CytochalasinB (Sigma) for 1 h or 10 μM Nocodazole (Sigma)for 20 min. Cells were then washed with prewarmed PBS, fixed for20 min with ice-cold methanol, permeabilized with a solution of PBSand 0.2% Triton X-100, washed again with PBS and saturated with PBS with 2% BSA.Cells were then incubated with the primary antibodies overnight at4°C (Table S2), washed with PBS three times 10 min,incubated with secondary antibodies and washed. Laser-scanning confocalmicroscopy was performed using a Zeiss LSM 710 laser scanning system.

Cell fractionation into cytosolic and microtubule fractions

An equivalent number of transfected cells was recovered in equal volumes ofwarmed lysis buffer: 1 mM MgCl2, 2 mM EGTA, 30%glycerol, 0.1% Triton X-100, 80 mM Pipes, pH 6.8, complemented withprotease inhibitors. After ultracentrifugation at 100 000 × g at21°C for 18 min, supernatants were collected as cytosolicfractions. The remaining pellets were washed and recovered as microtubulefractions by sonication in Ripa buffer (in a volume equal to that of cytosolicfractions). Samples were mixed with LDS buffer and equal volumes were loadedfor SDS-PAGE and analyzed by immunoblotting.

The methods were carried out in “accordance” with theapproved guidelines.