Báo cáo Y học: Formation of aberrant phosphotau fibrillar polymers in neural cultured cells doc

Formation of aberrant phosphotau fibrillar polymers in neural cultured cells Mar Pe ´ rez 1 ,Fe ´ lix Herna ´ ndez 1 , Alberto Go ´ mez-Ramos 1 , Mark Smith 2 , George Perry 2 and Jesu ´ s Avila 1 1 Centro de Biologı ´ a Molecular (CSIC/UAM), Facultad de Ciencias. Universidad Auto ´ noma de Madrid, Spain; 2 Institute of Pathology, Case Western Reserve University, Cleveland, OH, USA Here we show, for the first time, the in vitro formation of filamentous aggregates of phosphorylated tau protein in SH-SY5Y human neuroblastoma cells. The formation of such aberrant aggregates, similar to those occurring in vivo in Alzheimer’s disease and other tauopathies, requires okadaic acid, a phosphatase inhibitor , to increase the level of phos- phorylated tau, and hydroxynonenal, a product of oxidative stress that selectively adducts and modifies phosphorylated tau. Our findings suggest that both phosphorylation and oxidative modification a re r equired f or tau filament forma- tion. I mportantly, the in vitro formation of intracellular tau aggregates could be used as a model o f tau polymerization and facilitate the development of novel therapeutic approaches. Keywords: Alzheimer’s disease; tauopathies; oxidative stress; tau phosphorylation; aberrant aggregates. Tauopathies are a heterogeneous group of dementias sharing a common pathological hallmark, the presence of aberrant tau filaments or forms of tau [1,2]. Tau is a microtubule-associated protein [3,4] that in pathological situations, and in a hyperphosphorylated form [5–8] assembles into fibrillar polymers. The mechanism for that aberrant tau assembly has been widely analyzed by several groups, i ndicating that sulfated glycosaminoglycans or other anionic compounds could favour tau polymerization [9–11]. Another category of agents suggested to alter assembly are fatty acids t hat can facilitate a ggregation e ither dire ctly [12–15] and/or additionally through a reaction with the highly reactive products of lipid oxidation [16,17]. Ad di- tionally, proteolysis [18] and o ther tau modifications such as phosphorylation, glycation or oxidation [19–23] could p lay a role in the aberrant tau aggregation. The most studied mechanism of tau polymerization is phosphorylation, although under m any conditions, hyper- phosphorylated tau does not show a high c apacity f or in vitro polymerization as compared t o unmodified tau [24]. This represents a pivotal paradox as mutations linked t o familiar A lzheimer’s disease (AD, the mo st common tauopathy) like those f ound in presenilin-1 (PS1) a nd amyloid b protei n precursor ( APP), result in an i ncrease in both the level of phosphorylation of tau protein [ 25,26] and in its aggregation leading to neurofibrillary tangles. None- theless, it is clear that phosphorylated, but not unmodified tau, is able to polymerize in vitro in the presence of 4- hydroxynonenal (HNE), a naturally occurring p roduct of lipid peroxidation [17] that is increased in AD [27]. T o extend these latter studies, in this work we investigated the effect of HNE on tau in d ifferent phosphorylation status within neuroblastoma cells. As in cell f ree systems, t au phosphorylation is essential to HNE induced assembly. MATERIALS AND METHODS Materials Okadaic a cid ( OA) was purchased f rom Sigma. 4HNE wa s prepared as described previously [28]. P HF-1 antibody reacting with phosphotau [29] was a kind gift of P. Davies (Albert E instein College, Bronx, NY, USA); 7.51 a nd BR134 antibodies, reacting w ith t au protein, were a kind gift of C.M. Wischik (MRC, Cambridge, UK) [30]; a polyclonal a ntibody specific for the lysine-derived pyrrole adducts formed by HNE was used [27]. Alkaline phospha- tase was purchased from Roche. Gel electrophoresis and Western blot Cells were harvested i n chilled N aCl/P i , resuspended and homogenized in buffer c ontaining 50 m M Hepes, pH 7.4, 10 m M EDTA, 0.1% Triton X -100, 100 m M NaF, 0.1 m M sodium orthovanadate, 1 m M phenylmethanesulfonyl fluoride, 10 lgÆmL )1 leupeptin, 1 0 lgÆmL )1 pepstatin a nd 10 lgÆmL )1 aprotinin. Lysates were centrifuged at 10 000 g for 30 min at 4 °C a nd boiled f or 5 min in electrophoresis sample buffer. The amount of protein in the samples was quantitated by the BCA protein a ssay. SDS/PAGE was carried out using 10% gels, w hich were afterwards transferred to n itrocellulose to be tested with different antibodies. Immunoreactive proteins w ere visualized by Correspondence to J. Avila, Centro de Biologı ´ a Molecular (CSIC/ UAM), Facultad de Ciencias. Universidad Auto ´ noma de Madrid, Cantoblanco 28049, Madrid, Spain. Fax: + 34 91 3974499, Tel.: + 34 91 39 78440, E-mail: javila@cbm.uam.es Abbreviations: AD, Alzheimer’s disease; APP, amyloid precursor protein; DMEM, Dulbecco’s modified Eagle’s medium; HNE, 4-hydroxynonenal; NFT, neurofibrillary tangles; OA, okadaic acid; PHFs, paired helical filaments; PP1, protein phosphatase 1; PP2A, protein phosphatase 2A; PS1, presenilin-1; PSP, progressive supranuclear palsy. (Received 3 1 October 2 001, revised 1 1 January 20 02, accepted 18 January 2 002) Eur. J. Biochem. 269, 1484–1489 (2002) Ó FEBS 2002 chemiluminiscence detection (ECL kit from Pierce). Quan- titation of immunoreactivities was performed by densito- metric scanning. Cell culture SH-SY5Y neuroblastoma cells Human neuroblastoma SH-SY5Y cells [31] were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemen- tedwith10%fetalbovineserumand2m M glutamine plus 0,01% pyruvate i n a humidified atmosphere wit h 7 % CO 2 . The day before the experiment, the cells were subcultured, and a cell s uspension was placed into the w ells. A fter overnight incubation in growth medium, the SH-SY5Y cells were washed and incubated in DMEM w ithout fetal bovine serum c ontaining vehicle, 0.25 l M OA, 10 l M HNE or OA plus HNE for 45 min. Immunofluorescence analysis Cells plated on polylysine-coated coverslips w ere fi xed w ith 4% paraformaldehyde for 30 min. Dephosphorylation of phosphotau in fixed cultured cells using alkaline phospha- tase were carried out as described by Mattson et al. [32]. After t he incubation, the c overslips w ere w ashed with NaCl/P i supplemented with 0.1% Triton X-100 (NaCl/P i / Triton), for 10 min, then were incubated with 1 % fetal bovine serum in NaCl/P i /Triton for additional 1 0 min. Incubation with primary antibodies was carried out in NaCl/P i /Triton for 45 min at room temperature. Coverslips were rinsed three t imes with NaCl/P i /Triton and incuba ted for 30 min with Oregon green or Texas-Red conjugated secondary antibodies (1 : 400; Molecular Probes). F inally, cells were r insed with N aCl/P i /Triton a nd mounted i n Fluoromount. Coverslips were analyzed using a Zeiss epifluorescence microscope. Films were scanned in Filmscan 200 (EPSON), and images were processed in Adobe PHOTOSHOP 5.02 on a PC workstation. Isolation of PHF and tau filaments Brain samples, supplied by R. Ravid (Netherlands Brain Bank), from AD patients, were used as a source to isolate PHFs, by following the procedure of Greenberg and Davies [29]. To obtain filaments from SH-SY5Y cells, cells were homogen ized in buffer A (0.1 M Mes, pH 6.5, 0.5 m M MgCl 2 ,2m M EGTA, 0.5 M NaCl, 1 m M phenylmethanesulfonyl fluoride, 10 lgÆmL )1 leupeptin, 10 lgÆmL )1 pepstatin and 10 lgÆmL )1 aprotinin) by using a P otter homogenizer provided with a loosely fitting Teflon pestle. H omogenates were analyzed by direct adsorption of the samples to electron microscopy grids. Western blots studies, using Ab 7.51, homogenates from SH-SY5Y neuro- blastoma cells cultured in a P100 dish, were centrifuged at the highest speed of a Beckman a irfugue centrifuge (100 000 g) for 1 h, and the pelleted protein was tested. Electron microscopy and immunoelectronmicroscopy To t est for the presence of intracellular aggregates, untreat- ed or treated cells were fixed with 4% paraformaldehyde and 2 % g lutaraldehyde in cacodylate buffer for 60 min at 4 °C. SH-SY5Y cells were co llected a nd spun down at 1000 g for 5 min The pellet was postfixed i n 1% osmium tetroxide for 1 h and, afterwards, in 1 % u ranyl a cetate. After d ehydration with graded alcohols, the pellets were embedded in Epon and polymerized at 60 °C for 48 h. Ultrathin sections were observed by electron microscopy. To test for the presence of isolated filaments, s amples were placed on a carbon-coated grid fo r 2 min and then stained with 2% (w/v) uranyl acetate for 1 min. Transmis- sion electron microscopy was performed in a JEOL M odel 1200EX electron microscope operated at 100 kV. Electron micrographs were obtained at a magnification of 40 000 on Kodak SO-163 film. Immunoelectron m icroscop y was perform ed after adsorption of the s amples to electron microscopy grids and an incubation with the first antibody [(anti-HNE or anti-(tau BR134)], for 1 h a t room temperature, was performed. After extensive washing w ith NaCl/P i ,thegrids were incubated w ith the secondary antibody conju gated with 5-nm diame ter gold particles. Finally, the samples were negatively stained and observed, as described above. RESULTS Reaction of isolated PHF with an antibody raised against HNE In an earlier report, we showed increased a nd selective adduction of lipid peroxidation products such as HNE in association w ith n eurofibrillary tangles [27], the aberrant aggregates present in AD, and composed of bundles of paired helical filaments (PHF) [33]. To extend this, here w e determined that isolated PHF also c ontained p rotein-HNE conjugates (Fig. 1A) suggesting that HNE could play an in vivo role in the formation of tau aberrant aggregates. As a Fig. 1. Reaction of an antibody raised against HNE with isolated paired h elical filaments (PHF). PHF w ere isolated a s indicated in Methods and tested with Ab HNE (A) or with PHF-1 (B). T he result of that reaction is shown. B ar indicates 200 n m. Ó FEBS 2002 Tau aggregates in cultured cells (Eur. J. Biochem. 269) 1485 positive c ontrol, the r eaction of P HF with a tau antibody is shown in Fig. 1B. That HNE immunoreactivity is only seen on some regions of PHF may be due to HNE induced facilitation of tau–tau interaction. If so, HNE may be not easily accessible t o the antibody because it could be partially hidden in PHF structure. Another possibility could be t hat HNE modification may occur after PHF formation, being that modification not recognized by the antibody. Although, some HNE molecules could b e available to react with the antibody. This fact could also explain the relatively weak reaction of HNE antibodies with neurofibrillary tangles c ompared to t hat found in neuronal cytoplasm [ 27]. Tau in okadaic acid treated cells is in hyperphosphorylated form Treatment of SH-SY5Y human neuroblastoma cells with okadaic acid (OA), a phosphatase inh ibitor, results in the hyperphosphorylation of tau protein, as determined by its change in electrophoretic mobility a nd its reaction with Ab 7.51, an antibody that recognizes all tau isoforms inde- pendently of its phosphorylation status (Fig. 2, part I) or by the reaction with tau antibodies that specifically recognize phosphoepitopes, such as PHF-1 (Fig. 2, part III). This modified tau resembles that found in the brain of patients with different tauopathies [1]. T he OA-induced phosphory- lation of tau could be reversed by alkaline phosphatase treatment (data not shown). HNE treatment did not alter the level of tau phosphorylation found in OA treated cells as a similar pattern to that observed with OA alone (Fig. 2, part IIB) was observed in S H-SY5Y human neuroblastoma cells treated with OA/HNE (Fig. 2, part IID). Tau forms aberrant aggregates in neuroblastoma cells treated with okadaic acid and 4-hydroxynonenal Recently, we s howed, in a cell-free system [17] that phosphorylated tau, in the presence of H NE, polymerizes into fibrillar polymers. To test if, i n a similar way, tau could form aggregates in cultured cells, neuroblastoma cells were incubated in the absence (Fig. 3A), or the presence of O A (Fig. 3 B), HNE (Fig. 3C), or a mixture of OA/HNE (Fig. 3 D). As c learly shown in Fig. 3 cells treated with OA show an increase in PHF-1 immunoreactivity w ith a diffuse pattern (Fig. 3B) while in cells treated with OA/ HNE the pattern of PHF-1 immunoreactivity was clearly present in patches (Fig. 3D). Patches were present i n 9.2 ± 1.1% (n ¼ 4 i ndependent experiments) of the cell treated w ith OA/HNE. T hus, OA/HNE t reatment may result in the formation of aberrant aggregates (patches) distributed through the cytoplasm. These aggregates could be stained with antibodies raised against phosphotau (PHF-1). Notably, in thes e aggregates, tau phosphorylation is partially resistant to t he action of alkaline phosphatase (data not shown), pheno menon that has been also observed in cultured rat hippocampal n eurons [32]. T hese data suggest that tau is in a polymerized or aggregated f orm, similar to that observed i n t auopathies, such as AD , where tau phosphoepitopes are masked and dephosphorylation by AP is similarly restricted. Electron microscopy analysis of neu roblastoma cell sections also suggests the existence of aberrant filamentous aggregates in OA/HNE treated neuroblastoma cells (Fig. 4 ). These filamentous aggregates were not found in control, OA or HNE treated cells. Tau filaments are assembled in OA/HNE treated neuroblastoma cells The p revious results suggest that tau aggregates found in OA/HNE treated neuroblastoma cells could be composed Fig. 2. Okadaic acid (OA) treatment results in an increase of tau phosphorylation. S H -SY5Y neuroblastoma cells were treated for 45 min in the absence ( A), or presence of 0.25 l M OA (B), 10 l M HNE (C) or in th e presence o f both (D). Then, the cells were lysed and the presence of tau w as analyzed by gel electrophoresis and Western blot by using tau antibody 7.51 (I and II), or tau antib ody PHF-1 (III). (I) For 7.51 reaction, a decrease i n electrophoretic mobility was found in OA treated c ells. ( II) No differences in electrophoretic mobility w ere found f or samples treated with OA (B) or OA/HNE ( D). (III) For PHF-1 reaction, an increase in that reaction was found in the presence of OA (B, 4.55 ± 0.46-fold over control c ells) and in OA/HNE treated cells (D, 4.20 ± 0.86-fold over control c ells) compared to t hat found in the absence of treatment ( A, control c ells) or t he presence o f HNE ( 0. 80 ± 0.18-fold ove r control cells). Fig. 3. Formation of t au aggregates in neuroblastoma c ells. SH -SY5Y neuroblastoma cells were incubated in the absence (A), or presence of 0.25 l M OA (B), 10 l M HNE (C), or b oth (D). The presence of aggregates is indicated i n (D), after i mmuno fluorescence by using PHF-1 t au antibody. T he arro w shows th e aggregate, an d t he arrowhead the nucleus. Bar indicates 15 lm. 1486 M. Pe ´ rez et al. (Eur. J. Biochem. 269) Ó FEBS 2002 of fibrillar polymers. Thus, to confirm t he data, we tried to further c haracterize t hose fi laments from t he cell homogen- ate. Figure 5A–C) shows the presence in OA/HNE treate d cells of fibrillar polymers of 2–3 nm width, and, in some cases, wider 10 nm polymers were also found (not shown). These filaments were straight, and they did not presented a twisted s tructure. These filaments could b e stained with tau antibody BR134 (Fig. 5D–F) but a weak, if any, reaction with anti-HNE Ig was observed. These filaments were not found in control, OA or HNE treated cells. These results indicate that phosphotau protein could form aberrant filaments, in cultured cells, in the presence of a c ompound derived from oxidative stress. DISCUSSION Previous studies i n cell-free s ystems had suggested that phosphorylation and HNE binding act synergistically to promote tau aggregation [17]. In this study, we further extend those observations but in a context that mimics those conditions in which wild-type phosphotau forms polymers in tauopathies. We have also found here that HNE -adducts are associated with PHF th e component of neurofibrillary tangles ( NFT). HNE, a product of lipid peroxidation has been found associated in vivo with NFT [27] and it is able to modify in a way that results in the in vitro assembly of PHF like filaments [17]. Interestingly, it has been recently published that lipid peroxidation also precede s amyloid p laque formation [ 34] giving a strong support to a possible role of HNE in the formation of the aberrant structures found in AD. The second essential element for tau assembly is its phosphorylation. We have previously found that HNE reacts with normal t au and induces the A lz50 epitope in tau [35]. It is important that the a bility of HNE to create the Alz50 epitope not only is dependent on lysine residues of tau but also requires tau phosphorylation because neither methylated, recombinant, nor dephosphorylated tau reacts with HNE to create the Alz50 epit ope [35]. In this study, w e found that tau phosphorylation and HNE treatment of neuroblastoma cultured cells results in the assembly of tau into aberrant polym ers similar t o those found in human tauopathies. This provides the f oundation of a g ood m odel to test different compounds that could prevent abnormal tau aggregation. ThepolymersassembledinOA/HNEtreatedneuro- blastoma cells are partially resistant t o alkaline phospha- tase, a feature previously described [ 32] a nd that is also observed in tau filaments from some tauopathies. These polymers f rom neuroblastoma cells have mainly a diameter of 2–3 n m and are similar to t hose isolated from the brain Fig. 5. Presence of tau filaments in OA-plu-HNE t reated ne uroblasto- ma cells. E lectron microscopy o f negatively stained filaments (A,B ,C) and i mmunogold electron mic roscopy (D,E,F) with tau antibod y BR134. A secondary antibody conjugated with 5 n m diameter gold particle was u sed. 20–30 filaments w ere found per c arbon-coated grid loaded with protein obtained fro m OA/HN E t reated cells. No fi la- ments were observed in control, OA- or HNE-treated cells Samples were obtained as described in M aterials and m ethods. S cale bar rep- resents 100 nm. Fig. 4. Aberrant aggregates in OA ± H NE treated neuroblastoma cells. The presence of intracellular filamentous aggregates were observed in some cell sections after OA + HNE t reatment (see Materials a nd methods). Bar indic ates 200 nm. Ó FEBS 2002 Tau aggregates in cultured cells (Eur. J. Biochem. 269) 1487 of progressive supranuclear palsy (PSP) patients [36]. Nevertheless, wider polymers, similar to those described in other tauopathies a re also found. It is not known if 2 –3 nm polymers could be p recursors for the f ormation of wider (10 n m) polymers. Additionally, fi laments p resent in OA/HNE treated cells are not twisted s uggesting that some additional factors could be n ecessary to obtain twisted filaments [37]. Recently, it was described in non-neural cells that transfection with tau cDNA carrying some of t he muta- tions present in a tauopathy, frontotemporal d ementia linked t o chromosome 17 (FTDP-17), r esults in the expression of the mutated protein and in the f ormation of aberrant tau aggregates, indicating that these a ggregates couldbeassembledinculturedcells[38].However,inother tauopathies such a s Alzheimer’s disease tau mutations are not required f or the formation of aberrant aggregates [1] and only a post-translational modification, phosphoryla- tion, has been proposed to play a r ole in tau assembly [8,39]. This r ole h as bee n recently t ested [ 40] and, in agreement with our data, their results suggest that phos- phorylated tau has a high er c apacity f or self-assembly than unmodified tau. It is not well known how tau phosphorylation is promoted. It has been suggested that proteins such as beta amyloid [ 41] or presenilin 1 (in m utated form) [26,42] c ould induce tau phosphorylation through the activation of GSK3. On the other hand, some other tau protein k inases could be activated by other ways, such as by oxidative stress [43,44]. These kinases could play a role in tauopathies such as Alzheimer’s disease. T he residues modified b y those protein kinases in tau protein, c ould be dephosphorylated by the action of some okadaic acid-sensitive phosphatases such as PP2A or PP1 [45,46]. Thus, we have t reated neural cells with OA to increase cellular phosphorylated tau. A fact that was tested by t he use of an antibody (PHF-1) that recognized tau in phosphorylated form (also an increase in tau phosphorylation was found by testing with two other antibodies, AT8 and 12E8, that recognize other phospho- epitopes i n tau). H owever, tau phosphorylation is n ot sufficient for its aggregation. This work shows that a second element, HNE, is required for tau aggregation. HNE c an easily pass through neuronal c ompartments and bind t o tau protein [32]. If tau is phosphorylated, the reaction with HNE modifies i ts conformation [35] and promotes its assembly into fibrillar polymers [17] resembling NFTs [47]. Phosphorylation could facilitate a tau conform- ational c hange that may allow t he interaction of HNE with those tau regions mainly involved in polymer formation. One o f t hese regions is the third tubulin-binding motif present i n t au molecule [11]. Thus, HNE binding domain is inside the filament structure and an anti-HNE Ig may not be able to label the OA/HNE treated fi laments. An additional possibility could b e t hat HNE may suffer a modification after filament formation or t hat a twisted process would be necessary to expose the HNE epitope. In sum mary, we demonstrate here t hat intraneuronal hyperphosphorylated t au, i n the p resence of a natural peroxidation product, HNE, forms fibrillar polymers. This process likely resembles the mechanism responsible for the formation of aberrant tau aggregates present in tauopathies where both phosphorylation a nd lipid peroxidation are concurrent features of disease. ACKNOWLEDGEMENTS This work was supported by grants from Spanish CICYT, Comunidad de Madrid, Neuropharma a nd by an institutional grant from Fundacio ´ n R. Areces and by the National Institutes of Health (to GP). A predoctoral fellowship from Gob ierno Vasco was awarded to A. Go ´ mez-Ramos. The help of R. Cuadros and S. Soto-Largo i s acknowledged. Also, we a cknowledge to Dr J. J . Lucas for c ritical reading of t he manuscript. REFERENCES 1. Goedert, M., Spillantini, M. G. & Davies, S .W. (1998) Filamentous nerve cell inclusions in neurodegenerative diseases. Curr. Opin. N eurobiol. 8, 619–632. 2. Spillantini, M.G. & Goedert, M. (1998) T au protein pathology in neurodegenera tive diseases. Trends Neurosci. 21, 4 28–433. 3. Cleveland, D.W., Hwo, S .Y. & Kirschner, M.W. (1977) Physical and chemical prop erties of purifi ed tau fac tor and t he role of t au in microtubule assembly. J. Mol. Biol. 116, 2 27–247. 4.Weingarten,M.D.,Lockwood,A.H.,Hwo,S.Y.&Kirschner, M.W. (1975) A protein factor essential for microtubule assembly. Proc. Natl Acad. Sc i. USA 72 , 1858–1862. 5. Bramblett, G.T., Goedert, M., Jakes, R., Merrick, S.E., Trojan- owski, J.Q. & Lee, V.M. (1993) Abnormal tau phosphorylation at Ser396 in Alzheimer’s disease r ecapitulates development an d contributes to reduced microtubule b inding. Neuron 10, 1089– 1099. 6. Brion, J.P., Couck, A.M., Rob ertson, J., Loviny, T.L. & Ander- ton, B.H. (1993) Neurofilament monoclonal antibodies RT97 and 8D8 recognize different modified epitopes in p aired helical filament-tau in A lzheimer’s disease. J. Neurochem. 60, 1372–1382. 7.Goedert,M.,Jakes,R.,Crowther,R.A.,Six,J.,Lubke,U., Vandermeeren, M., Cras, P., Trojanowski, J.Q. & Lee, V.M. (1993) The abnormal phosphorylation of tau protein at Ser-202 in Alzheimer dise ase re capitulates p ho sphorylation during develop- ment. Proc. N atl Acad. Sc i. USA 90, 5066–5070. 8. Grundke, I.I., Iqbal, K., Quinlan, M., Tung, Y .C., Zaidi, M.S. & Wisniewski, H.M. (1986) Microtubule-associated protein tau. A component of Alzheimer paired helical filament s. J. Biol. Chem. 261, 6084–6089. 9. Goedert, M., Jakes, R., Spillantini , M.G., H asegawa, M., S mith, M.J. & Crowther, R.A. (1996 ) Assembly of microtubule-asso- ciated protein t au into Alzheimer-like filame nts in duce d by sulphated glycosaminoglycans. Nature 383, 550–553. 10.Kampers,T.,Friedhoff,P.,Biernat,J.,Mandelkow,E.M.& Mandelkow, E. (1996) RNA stimulates aggregation of micro- tubule-associated protein tau into Alzheimer-like paired helical filaments. FEBS Lett. 399, 344–349. 11. Pe ´ rez,M.,Valpuesta,J.M.,Medina,M.,MontejodeGarcini,E.& Avila, J. (1996) Polymerization of tau into filaments in t he presence of heparin: the minimal sequence required for tau–tau interacti on. J. Neurochem. 67 , 1183–1190. 12. Abraha, A., Ghoshal, N., Gamblin, T.C., Cryns, V., Berry, R.W., Kuret, J. & Binder, L.I. (2000) C-terminal inhibition of tau assembly in vitro and in Alzheimer’s disease. J. Cell Sci. 113, 3737–3745. 13. Gamb lin , T.C., Ki ng, M.E., Kuret, J., Berry, R.W. & Binder, L.I. (2000) Oxidative regulation of fatty acid-induced tau polymerization. Bi ochemistry 39, 1 4203–14210. 14. Wilson, D.M. & Binder, L.I. (1995) Polymerization of micro- tubule-associated protein tau under near- physiological conditions. J. Biol. Chem. 270, 2 4306–24314. 15. Wilson, D.M. & Binder, L.I. ( 1997) Free fatty acids stimulate the polymerization of tau and amyloid beta peptides: in vitro eviden ce for a common effector of pathogenesis in Alzheimer’s disease. Am. J Pathol. 150, 2181–2195. 1488 M. Pe ´ rez et al. (Eur. J. Biochem. 269) Ó FEBS 2002 16. Uchida, K. & S tadtman, E.R. (1 993) Covalent attach ment of 4 - hydroxynonenal to glyceroaldehyde-3 p hosphate d ehydrogenase. J. Biol. Chem. 268, 6388–6393. 17. Pe ´ rez, M., Cuadros, R., Smith, M.J., P erry, G. & Avila, J. ( 2000) Phosphorylated, but no native, tau protein assembles following reaction with the lipid product, 4-hyd roxy-2-nonenal. FEBS L ett. 486, 2 70–274. 18. Wischik, C.M., Novak, M., Edwards, P.C., Klug, A., T ichelaar, W. & Cro wther, R.A. (1988) Structural characterization of the core of the paired helical filament of Alzheimer disease. Proc. Natl Acad.Sci.USA85, 4884–4888. 19. Grundke-Iq bal, I., Iqbal, K., Tun g, Y.C., Quinlan, M ., Wisniewski, H.M. & Binder, L.I. (1986) Abnormal phosphor- ylation of t he micro tubule-assoc iated p rotein t au in Alzheimer cytoskeletal pathology. Proc. Natl. Acad. Sci. USA 83, 4913–4917. 20. Ledesma,M.D.,Perez,M.,Colaco,C.&Avila,J.(1998)Tau glycation i s i nvolved in aggregation o f the protein b ut not in t he formation of fi laments. Cell M ol. Biol. 44, 1111–1116. 21. Schweers, O., Mandelkow, E.M., Biernat, J. & Mandelkow, E. (1995) Oxidation of c ysteine-322 in the r epeat domain of m icro- tubule- associate d protein t au con trols the in vitro assembly of paired helical filaments. Proc. Natl Acad. Sci. USA 92, 8463–8467. 22. Troncoso, J.C ., Costello, A., Watson, A.L. Jr & Johnson, G.V. (1993) In vitro polymerization of oxidized tau into filaments. Brain Res. 613 , 313–316. 23. Yan, S.D., Chen, X., Schmidt, A.M., Brett, J., Godman, G., Z ou, Y.S.,Scott,C.W.,Caputo,C.,Frappier,T.&Smith,M.A.et al. (1994) Glycated tau protein in Alzheimer disease: a mechanism for induction of oxidant stress. Proc. Natl Acad. Sci. USA 91, 7 787– 7791. 24. Schneider, A., Biernat, J., vonBergen, M., Ma ndelkow, E. & Mandelkow, E.M. (1999) Phosphorylation that detaches tau protein from microtubules (Ser262, Ser214) also protects it against aggregation into Alzheimer paired h elical filaments. Biochem. USA 38, 3549–3558. 25. Alvarez, G., MunozMontano, J.R., Satrustegui, J., Avila, J., Bogonez, E. & D iazNido, J. (1999 ) L ithium protec ts cultured neurons again st beta-amyloid-induced neurodegeneration. F EBS Lett. 453, 260–264. 26. Pigino, G ., Pelsman, A., Mori, H . & Busciglio, J. ( 2001) Presenilin-1 mutations r educe cytoskeletal association, deregulate neurite growth, and p otentiate n euronal dystrophy and tau phosphorylation. J. Neurosci. 21, 834–842. 27. Sayre, L .M., Zelasko, D.A., Harris, P.L.R., Perry, G., Salomon, R.G. & Smith, M.A. (1997) 4-Hydroxynonenal-derived advanced lipid peroxidation end products are increased in Alzheimer’s dis- ease. J. Neurochem. 68 , 2092–2097. 28. Xu, G., L iu, Y. & Sayre, L.M. (1999) Independent s ynthesis, solution behavior, and studies on the mechanism of formation of primary amine-derived fluoropho re representing cross-linking of proteins by (E)-4-hydroxy-2-nonenal. J. Org. Chem. 64, 5732– 5745. 29. Greenberg, S.G. & Davies, P. (1990) A preparation of Alzheimer paired helical filaments that displays distinct tau proteins by polyacrylamide gel electrophoresis. Proc. Natl Acad. Sci. USA 87, 5827–5831. 30. Novak, M., Wischik, C.M., E dw ards, P., Pannell, R. & Milstein, C. (1989) C haracte risation of the first mon oclonal antibody against the pronase resistant core of the Alzheimer PHF. Prog. Clin. Biol R es. 317, 7 55–761. 31. Biedler, J.L., Ro ffler-Tarlov, S., Schaner, M. & Freedman, L.S. (1978) Multiple ne urotransmitter synthesis by hum an neuro- blastoma cell lines and clones. Cancer Res. 38, 3751–3757. 32. Mattson, M.P., Fu, W.M., Waeg, G. & U chida, K. (1997) 4-hydroxynonenal, a p roduct of lipid peroxidation, inhibits dephosphorylation of the microtubule-associated protein t au . Neuroreport 8, 2275–2281. 33. Kidd, M. (1963) Paired helical filaments in electron microscopy of Alzheimer’s disease. Nature 197, 1 92–193. 34. Pratico, D., Uryu, K ., Leight, S., Trojanoswki, J.Q. & Lee, V.M. (2001) Increased lipid peroxidation prec edes amyloid plaque for- mation in an animal model of Alzheimer amyloidosis. J. Neurosci . 21, 418 3–4187. 35. Takeda, A., Smith, M.A., Avila, J., Nunomura, A., Siedlak, S.L., Zhu, X., P erry, G . & Sayre, L.M. (2000 ) I n Alzheimer’s disease, heme oxygenase is coincident with Alz50, an epitope of tau induced by 4 -hydro xy-2-nonenal modification. J. Neurochem. 75, 1234–1241. 36. Pe ´ rez,M.,Valpuesta,J.M.,MontejodeGarcini,E.,Quintana,C., Arrasate, M., Lopez Carrascosa, J.L., Rabano, A., deYebenes, J.G. & Avila, J. (1998) Ferritin is associated with the aberrant tau filaments present in progressive supranuclear palsy. Am. J. Pathol. 152, 1 531–1539. 37. Arrasate, M., Perez, M., Valpuesta, J.M. & Avila, J. (1997) Role of glycosaminoglycans in determining the helicity of paired helical filaments. Am. J. Pathol. 151 , 1115–1122. 38. Vogelsberg-Ragaglia, V., Bruce, J., Richter-Landsberg, C., Zhang, B., H ong, M., T rojanowski, J.Q. & Lee, V.M Y. (2000) Distinct FTDP-17 Missense mutations in Tau produce Tau aggregates and other pahotlogical phenotypes in transfected CHO cells. Mol. Biol. Cell. 11, 4093–4014. 39. Kosik, K .S. (1992) Alzheimer’s disease: a cell biological perspec- tive. Science 256 , 780–783. 40. Alonso, A., Zaidi, T., N ovak, M., G rundke-Iqbal, I. & I qbal, K. (2001) H yperphosphorylation induces self-assembly of tau into tangles of paired h elical filaments/straigh t filaments. Proc. Natl Acad.Sci.USA98, 6923–6928. 41. Busciglio, J., Lorenzo, A., Yeh, J . & Yankner, B.A. (1995) b- Amyloid fibrils induce tau phosphorylation and loss o f microtu- bule bi nding. Neuron 14, 879–888. 42. Takashima, A., Murayama, M., Murayama, O., Kohno, T., Honda, T., Yasutake, K., N ihonmatsu, N., M ercken, M., Yam - aguchi, H., Sugihara, S. & Wolozin, B . (1998) P resenilin 1 associates with glycogen synthase kinase-3 beta and i ts substrate tau. Proc. N atl Acad. Sci . USA 95 , 9637–9641. 43. Perry, G., Roder, H., Nunomura, A., Takeda, A., Friedlich, A.L., Zhu, X.W., R aina , A.K., Ho lb rook, N., Siedlak, S.L., Harris, P.L.R. & Smith, M.A. (1999) Activation of neuronal extracellular receptor kinase (ERK) in A lzheimer disease links oxidative stress to abnormal phosphorylation. Neuroreport 10, 2411–2415. 44. Zhu, X., Rottkamp, C.A., Boux, H., Takeda, A., Perry, G. & Smith, M.A. (2000) Ac tivation of p38 kinase links ta u phos- phorylation, oxidativ e stress, and cell c ycle-related e vents in Alzheimer disease. J. Ne uropathol. Exp. Neurol. 59, 880 –888. 45. Goedert, M., Jakes, R., Qi, Z., Wang, J.H. & Cohen, P. (1995) Protein phosphatase 2A is th e major enzyme in brain t hat dephosphorylates tau protein phosphorylated by proline-directed protein kinases or cyclic AMP-Dependent protein kinase. J. Neurochem. 65 , 2804–2807. 46. Gong, C.X. , Shaikh, S., GrlundkeIqbal, I. & Iqbal, K. (1996) Inhibition of protein p ho sphatase-2B (calcineurin) activity towards Alzheimer abno rmally phosphorylated tau b y neuro- leptics. Brain R es. 741, 9 5–102. 47. Montine, T.J., Amarnath, V., Martin, M.E., Strittmatter, W.J. & Graham, D.G. (1996) E-4-hydroxy-2-nonenal is cytotoxic and cross-links cytoskeletal proteins in P19 neuroglial cultures. Am. J. Pathol. 148 , 89–93. Ó FEBS 2002 Tau aggregates in cultured cells (Eur. J. Biochem. 269) 1489 . studies on the mechanism of formation of primary amine-derived fluoropho re representing cross-linking of proteins by (E)-4-hydroxy-2-nonenal. J. Org. Chem mainly involved in polymer formation. One o f t hese regions is the third tubulin-binding motif present i n t au molecule [11]. Thus, HNE binding domain