- Email: [email protected]
Aggregation of tau protein by aluminum
77
Brain Research, 628 (1993) 77-84 © 1993 Elsevier Science Publishers B.V. All rights reserved 0006-8993/93/$06.00
BRES 19405
Aggregation of tau protein by aluminum Clay W Scott *, Ann Fieles, Linda A. Sygowski, Claudia B. Caputo Pharmacology Department, ICI Pharmaceuticals Group, ICI Americas, LW-2, Wilmington, DE 19897, USA (Accepted 22 June 1993)
Key words: Alzheimer's disease; Tau; Aluminum; Microtubule; Cytoskeleton; Deferoxamine
A l u m i n u m has been detected in Alzheimer neurofibrillary tangles, but the significance of its presence is unknown. The principal component of tangles is the paired helical filament (PHF), comprised of tau protein. We investigated whether aluminum could induce tau protein to form filaments or aggregate. W h e n 10/~M bovine tau or non-phosphorylated recombinant h u m a n tau was combined with 400 ~ M or more aluminum, tau protein appeared to aggregate, observed as a dose-dependent decrease in electrophoretic mobility on SDS-PAGE. Tau appeared as a smear above the region of the expected tau bands and, at higher aluminum doses, failed to enter the gel. A tau fragment encompassing the microtubule binding domains did not show decreased mobility in the presence of aluminum, but did form aggregates that failed to electrophorese. However no fibrillar structures were observed in the aluminum-treated tau samples when observed by electron microscopy. The effect of aluminum on tau mobility was reversed by incubating with 1 m M deferoxamine. In contrast, the morphology of P H F fibrils was unaffected by deferoxamine treatment and the characteristic abnormal mobility of PHF-tau was not reduced by deferoxamine. This suggests that aluminum is not, by itself, a significant factor in maintaining the assembly of PHF-tau as fibrils or in its abnormal mobility on SDS gels. A l u m i n u m treatment of 3T3 fibroblasts transfected with h u m a n tau resulted in toxicity, but did not change tau expression levels or induce tau aggregation. In conclusion, aluminum appears to induce isolated tau protein to aggregate in a phosphate-independent way, without the formation of fibrils. This effect was not observed when tau-transfected cells were treated with toxic doses of aluminum.
INTRODUCTION Aluminum is a neurotoxin that induces the formation of neurofibrillary tangles in vivo and in cultured cells t. These tangles appear to contain aggregates or bundles of hyperphosphorylated neurofilaments 3'35. Purified neurofilaments aggregate in the presence of aluminum and show reduced electrophoretic mobility on SDS-PAGE when incubated with aluminum 22'27'33. In addition, neurofilament phosphopeptides adopt /3pleated sheet conformations in the presence of aluminum 15. The neurofibrillary tangles of Alzheimer's disease are comprised of paired helical filaments (PHFs), whose morphology differs from the straight fibrils present in aluminum-induced tangles 1. Alzheimer PHFs are comprised of tau protein, rather than neurofilaments 37. Nevertheless, many of the characteristics of aluminumtreated neurofilaments are similar to those of tau protein in Alzheimer's disease. For instance, PHFs contain hyperphosphorylated tau in an aggregated (fibril-
* Corresponding author. Fax: (1) (302) 886-2766.
lar) state 21 and show extensive /3-pleated sheet structure 17. Tau extracted from PHFs also shows decreased electrophoretic mobility on SDS-PAGE 21. Aluminum has been detected in Alzheimer tangles ~4. The aluminum chelator, deferoxamine, has been reported to have therapeutic effects on Alzheimer dementia and to release aluminum from body stores in Alzheimer patients t 1. Although it is not known whether the deferoxamine treatment releases aluminum from tangles, these results imply that aluminum may play a role in Alzheimer dementia. PHFs are likely to be a major contributor to Alzheimer dementia 4'6. Therefore any effect of aluminum to induce or stabilize PHFs would be expected to worsen dementia. The significance of the presence of aluminum in Alzheimer tangles is controversial. Its presence suggests the possibility that aluminum may play a role in initiating the formation of PHFs 28, analogous to the role proposed for aluminum in Alzheimer plaque formation 5. Aluminum is known to bind to phosphoryl groups on proteins 2. Therefore a potential mechanism for this initiation of P H F formation by aluminum is the binding of aluminum to the phosphate groups on tau, inducing it to polymerize. The feasibility of this possi-
78 bility is suggested by the observation that phosphorylated neurofilaments are induced to aggregate by aluminum and adopt a /3-pleated sheet conformation ~s'22. However, another possibility is that aluminum binds to PHF components, before or after their assembly into fibrils, without inducing their formation or stabilizing them. We performed a series of experiments to determine whether aluminum may play a role in the assembly of tau into PHFs. In particular we investigated whether aluminum may be involved in causing tau to aggregate or form fibrils or to stabilize the P H F structure. We treated non-phosphorylated recombinant human tau, phosphorylated bovine tau, hyperphosphorylated PHFtau and tau transfected fibroblasts with aluminum under conditions which induce neurofilament aggregation. PHF-tau and the aluminum-treated tau samples were incubated with deferoxamine to evaluate the reversible nature of the aggregates. The aggregation and fibril assembly properties of these samples were evaluated using electron microscopy and SDS-PAGE. MATERIALS
AND METHODS
Tau proteins Bovine tau was isolated as described 29. The shortest isoform of human tau (7352), and a tau fragment (z264), which encompasses amino acids 264-386 of the longest human tau isoform, were expressed in E. coli and purified as previously described 3°'31. SDSsoluble PHFs, referred to as A68 PHFs 21, were prepared from Alzheimer cortical tissue by a modification 9 of the procedure of Ksiezak-Reding et al. 19. Cell cultures 3T3 fibroblasts were stably transfected with cDNA for the shortest isoform of human tau (~-352) as previously described 23. Cells were plated in 100 mm Nunc culture dishes or onto glass coverslips and grown in D M E M (Gibco) supplemented with 10% fetal bovine serum. From 1 to 10 mM aluminum lactate was added to the medium for 1-3 days. Viability of coverslip cultures was assessed using the Live/Dead viability/cytotoxicity assay kit (Molecular Probes, Inc.). Cell extracts were prepared as previously described 34. Protein content was measured by the method of Lowry et al. 24 after TCA precipitation of extracted protein. Tau content was determined by ELISA7 using antibody tau-14, diluted 1:6,000 and purified recombinant human 7352 at 0.3-300 nM to generate the standard curve. Monoclonal antibody z-14, which recognizes amino acids 141-178 of ~-441TM was generously provided by Dr. Virginia Lee. Other methods and reagents SDS-gel electrophoresis was performed on isolated proteins and on cell extracts and analyzed by Western blot analysis 34 and PROBLUE staining of proteins (Enprotech). Electron microscopy of protein samples that were negatively stained with 2% uranyl acetate and PHF samples negatively stained with 1% lithium phosphotungstic acid was performed as previously described 8. Deferoxamine (Desferal, CIBA Pharmaceutical Company) was hydrated in water prior to use. Aluminum perchlorate, aluminum lactate, aluminum sulfate, chromium chloride, ferric chloride and magnesium chloride were of reagent grade and were prepared as stock solutions immediately prior to use.
RESULTS
Bovine tau was incubated with increasing concentrations of aluminum perchlorate and then electrophoresed using reducing SDS-PAGE. As shown in Fig. 1, bovine tau protein migrated as three to five bands. Samples of bovine tau that were incubated with aluminum showed decreased electrophoretic mobility. The typical tau bands disappeared and were replaced with a smear of protein covering the 50-100 kDa range. In some samples a faint band at 130 kDa was also observed. This pattern suggests that aluminum induced tau conformations that were stable to SDS and boiling and appeared as slowly migrating species on reducing SDS gels. This effect was dependent on both the aluminum and tau concentrations. With 10 /xM bovine tau, the lowest effective dose of aluminum that caused a mobility shift was 200 /~M. At higher aluminum concentrations, further decreases in tau mobility were seen. At aluminum concentrations above 800 /~M, tau did not enter the gel matrix, presumably due to the formation of SDS-insoluble aggregrates. These effects are similar to those reported with aluminum and neurofilaments 22. Incubating the aluminum-treated tau samples with deferoxamine reversed the effect of aluminum on tau mobility, causing tau to migrate similar to the untreated control sample (Fig. 1B). Presum-
A
B 1 2
3
4
5 6
7
1
2
3
4
5
6
66-4.1--
22-14-Fig. 1. Effect of aluminum on the electrophoretic mobility of bovine tau. A: bovine tau (10/~M) was incubated alone (lane 2) or with 0.2 mM (3), 0.4 mM (4), 0.6 mM (5), 0.8 mM (6), or 1 mM aluminum perchlorate (7) at pH 6.7 for 30 min at 37°C. The samples were heated in sample buffer (62.5 mM Tris, pH 6.9, 1% SDS, 10% glycerol, 20 mM DTT, 0.005% Bromphenol blue), electrophoresed on a 10-20% gradient polyacrylamide gel and the gel stained with PRO-BLUE (Enprotech). Lane 1 contains molecular weight standards. B: bovine tau (10/xM) was incubated alone (lanes 1, 4) or with 0.4 mM (lanes 2, 5) or 1.0 mM aluminum perchlorate (lanes 3, 6) at pH 6.7 for 30 min at 37°C. Some samples were then incubated with 1 mM deferoxamine for 30 min (lanes 4-6). Samples were then heated in sample buffer and analyzed as described for A.
79
B
A
1
2
34
66
7
1
8
2 3 4 6 6
97-U-4S-66--
31--
43--
22-14--
31--
6--
14
g
Fig. 2. Effect of aluminum on the electrophoretic mobility of recombinant tau and a recombinant tau fragment. A: r352 (10 p.M) was incubated alone (lane 2) or with 0.2 mM (3), 0.4 mM (4), 0.6 mM (5), 0.8 mM (6), 1 mM (7) or 3 mM aluminum perchlorate (8) at pH 6.7 for 30 min at 37°C. Samples were analyzed as described in Fig. 1A. B: r264 (10 p~M) was incubated alone (lane 2) or with 0.2 mM (3), 0.5 mM (4), 1 mM (5) or 5 mM aluminum perchlorate (6) at pH 6.7 for 30 min at 37°C. Samples were heated in sample buffer, electrophoresed on a 10-20% gradient gel using Tris/Tricine buffer system (Enprotech), and stained with PRO-BLUE.
ably, deferoxamine was able to remove the aluminum which, when bound to tau, caused the change in tau mobility. Recombinant human tau (r352) was incubated with increasing concentrations of aluminum perchlorate and analyzed by SDS gel electrophoresis. As with the bovine tau samples, the aluminum-treated recombinant human tau samples ran as a smear of protein covering the 50-80 kDa range (Fig. 2A). A tau band at approximately 130 kDa was also observed. At aluminum concentrations greater than 1 mM tau formed aggregates that did not enter the gel matrix. Since recombinant tau is not phosphorylated 32, this result suggests that aluminum binds to the amino acid side chains of tau rather than to phosphoryl groups to elicit its effect on tau mobility. This effect was not dependent on the aluminum salt, as aluminum lactate and aluminum sulfate were also capable of inducing aggregation of bovine and recombinant tau. Ferric chloride did not cause a shift in tau mobility, although at 10 mM it did appear to create tau aggregates that did not enter the gel (data not shown). Ten mM magnesium chloride had no effect on tau mobility. Experiments were performed to see if aluminum could induce the aggregation of r264, a recombinant tau fragment. This fragment contains the microtubule binding domains of tau and encompasses the region of tau that forms the protease-resistent core of SDS-insoluble PHFs 16'29. Concentrations of aluminum that caused the shift in electrophoretic mobility of bovine
tau and recombinant tau to the 50-80 kDa range had no effect on the electrophoretic mobility of the tau fragment (Fig. 2B). However, at aluminum concentrations of 1 mM or greater the tau fragment failed to enter the gel matrix, presumably due to its aggregation into SDS-insoluble structures. Thus, the small shift in gel migration to 50-80 kDa was specific to full length taus and required protein sequence outside of the microtubule binding domain. The formation of large aggregates that do not electrophorese was a property of both full length taus and the tau fragment. A68 PHFs were isolated and experiments performed to see whether aluminum or deferoxamine affects the aggregation state of these fibrils. A68 PHFs are soluble in SDS, and the proteins that comprise these filaments can be separated by SDS gel electrophoresis 2~. The major component of PHFs is hyperphosphorylated tau protein 2t. PHF-tau shows reduced mobility on SDS gels, probably as a consequence of its abnormal phosphorylation state ~3. To determine whether aluminum may also contribute to the reduced mobility of PHF-tau, samples of A68 PHFs were treated with deferoxamine and then analyzed by SDS-PAGE. No increased mobility of PHF-tau was observed using 1-20 times the concentration of deferoxamine that was effective in reversing the aluminum induced mobility change with bovine and recombinant tau (Fig. 3). This suggests that either aluminum is not responsible for the reduced mobility of PHF-tau or aluminum (if it is bound to PHF-tau) has a much higher affinity for PHF-tau than recombinant or bovine tau, preventing its removal by deferoxamine. Interestingly, PHF-tau was similar to the other tau preparations in terms of its susceptibility to forming aggregates with aluminum that were stable to SDS treatment (Fig. 3). This effect was observed with both isolated A68 PHFs and SDS-solubilized PHF-tau. Samples of bovine tau, r352 and r264 that showed aluminum-induced mobility changes were examined by negative stain electron microscopy. No fibrillar structures were observed in these aluminum-treated samples (Fig. 4). Both aluminum and deferoxamine-treated A68 PHFs were also examined by electron microscopy. Neither treatment altered the number of fibrils present in the A68 P H F samples, or the morphology of the PHFs (data not shown). It is possible that a nucleating agent is required for aluminum-tau polymerization. Such an agent would not be present in studies using purified tau protein, but may be present in intact cells. Therefore, experiments were performed to see whether aluminum could induce the formation of tau fibrils in tau-transfected 3T3 fibroblasts. The morphology of tau-transfected 3T3 cells
80 is altered, with cells expressing ~-352 appearing more flattened than non-transfected fibroblasts z3. Aluminum lactate treatment produced a dose-dependent toxic effect on the transfected cells. The number of cells that stayed attached to the plate decreased as a function of aluminum concentration (Fig. 5), and the amount of protein extracted per culture also decreased (Table I). The morphology of the cells that remained attached also changed upon aluminum treatment (Fig. 5), with thin processes becoming prominent. These processes appear similar to those that are seen on 3T3 fibroblasts expressing other tau isoforms 23. The cells exhibiting these processes were still viable, as assessed by the L i v e / D e a d assay (Fig. 5). Toxicity was apparent after two days of treatment of low density cultures with 4 mM aluminum, while cells cultured at a higher density appeared slightly less sensitive to the treatment. Although aluminum treatment resulted in a decrease in the number of attached, viable cells and the total protein content of the cell extract, it did not cause a decrease in the level of tau expression based on percentage of total protein (Table I). In fact, at the higher aluminum doses tau expression accounted for a
1
2
3
4
97~ 68-3--
14-Fig. 3. Western blot showing the effects of aluminum and deferoxamine on the mobility of PHF-tau. Lane 1 contains bovine tau. Aliquots of A68 PHFs were incubated either alone (lane 2) or with 10 mM desferoxamine (lane 3) or 1 mM aluminum perchlorate (lane 4) under the conditions described in Fig. 1. The samples were electrophoresed then transferred to nitrocellulose and probed with antibody ~-14. PHF-tau was used at a concentration similar to bovine tau and recombinant tau, based on relative immunostaining intensities.
TABLE 1
Effect of aluminum lactate treatment of "r-transfected 3T3 cells on protein expression Cells were cultured at two densities and treated with aluminum for 3 days. Cell extracts were analyzed for total protein content and tau protein content as described in Materials and Methods.
Cell density
Aluminum concerttration (mM)
Total protein Ozg / lO0 mm dish)
Tau protein (% of total protein)
(n)
1×
0 2 4 6 8
130 121 79 42 18
3.9±1.6 2.8±0.8 1.9±0.5 2.1±0.5 4.8±1.4
6 6 6 5 4
0.5 ×
0 2 4 6 8
105 99 42 24 19
2.7±0.6 3.5±0.9 1.5±0.7 3.7±0.8 4.4±1.1
6 6 3 4 2
larger portion of the total protein. Aliquots of each extract, which contained equivalent protein contents and tau contents (based on ELISA), were applied to gels and electrophoresed to assess the state of aggregation of the expressed tau protein. No change in the electrophoretic mobility of tau was detected in the aluminum-treated cell extracts, compared to the nontreated cell extracts, when extract samples containing either 5 p.g (Fig. 6) or 10/xg protein (data not shown) were electrophoresed. The distribution of tau between the two bands, which represents different phosphorylation states of 734, was not affected by aluminum treatment. DISCUSSION It has been suggested that aluminum binds to the negatively charged phosphate groups on phosphoproteins such as neurofilaments, and thereby induces crosslinking and aggregation12'2% In this study we evaluated several different tau proteins which are phosphorylated to different extents, to determine whether the phosphorylation state altered the sensitivity of tau to form aggregates or fibrils. Recombinant human tau isoform ~-352 and tau fragment r264 are isolated as nonphosphorylated proteins 31 while bovine tau contains at least three phosphate groups 29. Hyperphosphorylated PHF-tau isolated from A68 PHFs contains up to eight phosphate groups 2° and human tau expressed in 3T3 fibroblasts migrates as two bands on gels, reflecting heterogenous phosphorylation states 3a. Aluminum caused the isolated tau proteins to aggregate, as defined by a decrease in migration on SDSP A G E and a loss of protein entering the gel matrix.
81 These effects are analogous to the aluminum effects on neurofilaments and occur with similar aluminum concentrations 22'27. The aluminum-induced tau aggregates were not fibrillar or PHF-like in morphology, but rather appeared as amorphous aggregates. Nonphosphorylated r352 and bovine tau were equally sensitive to aluminum, indicating that phosphorylation of tau at physiological sites does not enhance the aluminum effect. The recombinant fragment r264 formed aggregates in the presence of aluminum that failed to enter the gel matrix, but did not show the decreased mobility seen with intact tau. Presumably, regions outside of the microtubule binding domains are required to achieve this conformation. It cannot be discerned from these data whether this conformation reflects an intermolecular aggregation of tau monomers or an intramolecular conformation that migrates anomalously on SDS gels. Aluminum was toxic to fibroblasts that expressed tau. However tau did not aggregate in these cells, eliminating it as a mechanism for the observed toxicity.
In contrast, neurofilaments do aggregate in aluminumtreated cells =. The transfected fibroblasts treated with the highest doses of aluminum showed an increase in tau content. The increased tau content was probably not responsible for the toxicity, as cell death was observed with concentrations of aluminum below that which caused an increase in tau content. Furthermore, an aluminum-induced increase in tau immunoreactivity has been observed in human neuroblastoma ceils, without an effect on cell viability 26. Other mechanisms of aluminum toxicity that may have been responsible for the toxicity we observed include interference of aluminum with D N A function 1° or with magnesium-mediated biological processes, as aluminum and magnesium have similar ionic radii and can compete in various biological systems 35. Aluminum causes both tau and neurofilament monomers in soluble form to aggregate into non-fibrillar material. Aluminum also appears to induce neurofilaments in fibril form to bundle, forming tangles z2'36,
Fig. 4. Electron micrographs of PHFs (A) and aluminum treated tau preparations, r352 (B), r264 (C) and bovine tau (D) were incubated with 1 mM aluminum perchlorate for 30 rain at 37°C and then placed on carbon-formvar coated grids, stained with 1% lithium phosphotungstic acid and examined by electron microscopy.Magnification, 74,250 ×.
82
~
i¸
~ . . . .
e
~ ~
~
-
¸
Fig. 5. Phase contrast micrographs of 3T3 cells transfected with h u m a n r352 D N A and cultured in the absence (A), or presence of aluminum lactate at 2 rnM (B), 4 m M (C), 6 m M (D), or 8 m M (E). Arrowheads indicate dead cells as determined by L i v e / D e a d assay of the same field. No cells were dead in B, one cell was dead in A, C, and D; most of the cells were dead in E.
83 1
2
3
4
5
6
7
8
9
10
11
SSm 43--
2w
Fig. 6. Western blot of extracts of the cell cultures in Fig. 5 are shown. Aliquots of extracts, containing 5 /zg of protein, were electrophoresed, transferred to nitrocellulose and probed with antibody tau-14 for tau immunoreactivity. Lane 1 contains purified human ~-352 and ~-441 expressed in E. coll. Lanes 2-6 contain extracts of cells cultured at an initial plating density that was twice that of the cultures used for the extracts in lanes 7-11. Cultures were untreated (lanes 2, 7) or treated with aluminum at 2 mM (lanes 3, 8), 4 mM (lanes 4, 9), 6 mM (5, 10), or 8 mM (6, 11).
but does not induce similar tau-containing tangles to form under the conditions used in this study. Since neurofilament monomers normally assemble into fibrils, the effect of aluminum in vivo is to cause aggregation of the assembled fibrils of neurofilaments into large aggregates or tangles. However since tau does not normally form fibrils, aluminum treatment does not produce aggregates or tangles of tau fibrils nor does it act as a nucleating agent under the conditions tested to induce tau to form fibrils. The lack of effect of deferoxamine treatment on P H F morphology and electrophoretic mobility of PHF-tau suggests that aluminum is not required to maintain the P H F structure, nor is it responsible for the abnormal mobility of PHF-tau. Acknowledgments. The authors wish to acknowledge the technical assistance of Hope Koenigbauer, William Brunner and Irene Evangelista Sobel. The authors thank Dr. Claude Wischik and the Cambridge Brain Bank for providing Alzheimer cortical brain tissue and Dr. Michel Goedert for supplying the tau clones.
3 Bizzi, A. and Gambetti, P., Phosphorylation of neurofilaments is altered in aluminum intoxication, Acta Neuropathol., 71 (1986) 154-158. 4 Braak, H. and Braak, E., Neuropathological stageing of Alzheimer-related changes, Acta Neuropathol., 82 (1991) 239-259. 5 Candy, J.M., Oakley, A.E., Klinowski, J., Carpenter, T.A., Perry, R.H., Atack, J.R., Perry, E.K., Blessed, G., Fairbairn, A. and Edwardson, J.A., Aluminosilicates and senile plaque formation in Alzheimer's disease, Lancet, 2 (1986) 354-357. 6 Caputo, C.B. and Salama, A.I., The amyloid proteins of Alzheimer's disease as potential targets for drug therapy, Neurobiol. Aging, 10 (1989) 451-461. 7 Caputo, C.B., Wischik, C., Novak, M., Scott, C.W, Brunner, W.F., Montejo de Garcini, E., Lo, M.M.S., Norris, T.E. and Salama, A.I., Immunological characterization of the region of tau protein that is bound to Alzheimer paired helical filaments. Neurobiol. Aging, 13 (1992) 267-274. 8 Caputo, C.B., Sygowski, L.A., Scott, C.W and Sobel, I.R.E., Role of tau in the polymerization of peptides from /3-amyloid precursor protein, Brain Res., 597 (1992) 227-232. 9 Caputo, C.B., Hayward, C., Brunner, W.F., Scott, C.W and Sygowski, L.A., Association of the C-terminus of fl-amyloid precursor protein with a tau protein fragment in Alzheimer PHF material, Biochem. Biophys. Res. Commun., 185 (1992) 1043-1040. 10 Crapper McLachlan, D.R., Lukiw, W.J. and Kruck, T.P.A., New evidence for an active role of aluminum in Alzheimer's disease, Can. J. Neurol. Sci., 16 (1989) 490-497. 11 Crapper McLachlan, D.R., Dalton, A.J., Kruck, T.P.A., Bell, M.Y., Smith, W.L., Kalow, W. and Andrews, D.F., Intramuscular deferoxamine in patients with Alzheimer's disease, Lancet, 337 (1991) 1304-1308. 12 Ganrot, P.-O., Biochemistry and metabolism of AICI 3 and similar ions: a review, Environ. Health Perspect., 65 (1986) 363-441. 13 Goedert, M., Spillantini, M.G., Cairns, N.J. and Crowther, R.A., Tau proteins of Alzheimer paired helical filaments: abnormal phosphorylation of all six brain isoforms, Neuron, 8 (1992) 159168. 14 Good, P.F., Perl, D.P., Bierer, L.M. and Schmeidler, J., Selective accumulation of aluminum and iron in the neurofibrillary tangles of Alzheimer's disease: a laser microprobe (LAMMA) study, Ann. Neurol., 31 (1992) 286-292. 15 Hollosi, M., Urge, L., Perczel, A., Kajtar, J., Teplan, I., Otvos, L. and Fasman, G.D., Metal ion-induced conformational changes of phosphorylated fragments of human neurofilament (NF-M) protein, J. Mol. Biol., 223 (1992) 673-682. 16 Jakes, R., Novak, M., Davison, M. and Wischik, C.M., Identification of 3- and 4-repeat tau isoforms within the PHF in Alzheimer's disease, EMBO J., 10 (1991) 2725-2729. 17 K.irschner, D.A., Abraham, C. and Selkoe, D.J., X-ray diffraction from intraneuronal paired helical filaments and extraneuronal amyloid fibers in Alzheimer disease indicates cross-/3 conformation, Proc. Natl. Acad. Sci. USA, 83 (1986) 503-507. 18 Kosik, K.S., Orecchio, L.D., Binder, L., Trojanowski, J.Q., Lee, V.M.-Y. and Lee, G., Epitopes that span the tau molecule are shared with paired helical filaments, Neuron, 1 (1988) 817-825. 19 Ksiezak-Reding, H., Binder, L.I. and Yen, S.-H., Alzheimer disease proteins (A68) share epitopes with tau but show distinct biochemical properties, J. Neurosci. Res., 25 (1990) 420-430. 20 Ksiezak-Reding, H., Liu, W.-K. and Yen, S.-H., Phosphate analysis and dephosphorylation of modified tau associated with paired helical filaments, Brain Res., 597 (1992) 209-219. 21 Lee, V.M.-Y., Balin, B.J., Otvos, L. and Trojanowski, J.Q., A68 a major subunit of paired helical filaments and derivatized forms of normal tau, Science, 25 (1991) 675-678. 22 Leterrier, J.F., Langui, D., Probst, A. and Ulrich, J., A molecular mechanism for the induction of neurofilament bundling by aluminum ions, J. Neurochem., 58 (1992) 2060-2070. 23 Lo, M.M.S., Fieles, A.W., Norris, T.E., Dargis, P.G., Caputo, C.B., Scott, C.W, Lee, V.M.-Y. and Goedert, M., Human tau isoforms confer distinct morphological and functional properties to stably transfected fibroblasts, Mol. Brain Res., 20 (1993) in Dress. -
REFERENCES 1 Bertholf, R.L., Aluminum and Alzheimer's disease: Perspectives for a cytoskeletal mechanism, Crit. Rev. Clin. Lab. Sci., 25 (1987) 195-210. 2 Birchali, J.D. and Chappell, J.S., The chemistry of aluminum and silicon in relation to Alzheimer's disease, Clin. Chem., 34 (1988) 265-267.
-
84 24 Lowry, O.H., Rosebrough, N.J., Farr, A,L., and Randall, R.J., Protein measurement with the Folin phenol reagent, J. Biol. Chem., 193 (1951) 265-27. 25 Macdonald, T.L. and Martin, F.B., Aluminum ion in biological systems, Trends Biol. Sci., 13 (1988) 15-19. 26 Mesco, E.R., Kachen, C. and Timiras, P.S., Effects of aluminum on tau proteins in human neuroblastoma cells, Mol. Chem. Neuropathol., 14 (1991) 199-212. 27 Nixon, R.A., Clarke, J.F., Logvinenko, K.B., Tan, M.K.H., Hoult, M. and Grynspan, F., Aluminum inhibits calpain-mediated proteolysis and induces human neurofilament proteins to form protease-resistent high molecular weight complexes, J. Neurochem., 55 (1990) 1950-1959. 28 Perl, D.P. and Pendlebury, W.W., Aluminum neurotoxicity - potential role in the pathogenesis of neurofibrillary tangle formation, Can. J. Neurol. Sci., 13 (1986) 441-445. 29 Poulter, L., Barratt, D., Scott, C.W and Caputo, C.B., Locations and immunoreactivities of phophorylation sites on bovine and porcine tau proteins and a PHF-tau fragment, J. Biol. Chem., 268 (1993) 9636-9644. 30 Scott, C.W, Blowers, D.P., Barth, P.T., Lo, M.M.S., Salama, A.I. and Caputo, C.B., Differences in the abilities of human tau isoforms to promote microtubule assembly, J. Neurosci. Res., 30 (1991) 154-162. 31 Scott, C.W, Klika, A.B., Lo, M.M.S., Norris, T.E. and Caputo, C.B., Tau protein induces bundling of microtubules in vitro:
32
33
34
35
36
37
comparison of different tau isoforms and a tau protein fragment, J. Neurosci. Res., 33 (1992) 19-29. Scott, C.W, Spreen, R.C., Herman, J.L., Chow, F.P., Davison, M.D., Young, J. and Caputo, C.B., Phosphorylation of recombinant tau by cAMP-dependent protein kinase, J. Biol. Chem., 268 (1993) 1166-1173. Shea, T.B., Beerman, M.L. and Nixon, R.A., Aluminum alters the electrophoretic properties of neurofilament proteins: role of phosphorylation state, J. Neuroehem., 58 (1992) 542-547. Sygowski, L.A., Fieles, A.W., Lo, M.M.S., Scott, C.W and Caputo, C.B., Phosphorylation of tau protein in tau-transfected 3T3 cells, Mol. Brain Res., 20 (1993) in press. Troncosco, J.C., Sternberger, N.H., Sternberger, L.A., Hoffman, P.N. and Price, D.L., Immunocytochemical studies of neurofilament antigens in the neurofibrillary pathology induced by aluminum, Brain Res., 364 (1986) 295-300. Troncoso, J.C., March, J.L., Haner, M. and Aebi, U., Effect of aluminum and other multivalent cations on neurofilaments in vitro: An electron microscopic study, J. Struct. Biol., 103 (1990) 2-12. Wischik, C.M., Novak, M., Thogersen, H.C., Edwards, P.C., Runswich, M.F., Jakes, R., Milstein, C., Roth, M. and Klug, A., Isolation of a fragment of tau derived from the core of the paired helical filament of Alzheimer disease, Proc. Natl. Acad. Sci. USA 85 (1988) 4506-4510.
Brain Research, 628 (1993) 77-84 © 1993 Elsevier Science Publishers B.V. All rights reserved 0006-8993/93/$06.00
BRES 19405
Aggregation of tau protein by aluminum Clay W Scott *, Ann Fieles, Linda A. Sygowski, Claudia B. Caputo Pharmacology Department, ICI Pharmaceuticals Group, ICI Americas, LW-2, Wilmington, DE 19897, USA (Accepted 22 June 1993)
Key words: Alzheimer's disease; Tau; Aluminum; Microtubule; Cytoskeleton; Deferoxamine
A l u m i n u m has been detected in Alzheimer neurofibrillary tangles, but the significance of its presence is unknown. The principal component of tangles is the paired helical filament (PHF), comprised of tau protein. We investigated whether aluminum could induce tau protein to form filaments or aggregate. W h e n 10/~M bovine tau or non-phosphorylated recombinant h u m a n tau was combined with 400 ~ M or more aluminum, tau protein appeared to aggregate, observed as a dose-dependent decrease in electrophoretic mobility on SDS-PAGE. Tau appeared as a smear above the region of the expected tau bands and, at higher aluminum doses, failed to enter the gel. A tau fragment encompassing the microtubule binding domains did not show decreased mobility in the presence of aluminum, but did form aggregates that failed to electrophorese. However no fibrillar structures were observed in the aluminum-treated tau samples when observed by electron microscopy. The effect of aluminum on tau mobility was reversed by incubating with 1 m M deferoxamine. In contrast, the morphology of P H F fibrils was unaffected by deferoxamine treatment and the characteristic abnormal mobility of PHF-tau was not reduced by deferoxamine. This suggests that aluminum is not, by itself, a significant factor in maintaining the assembly of PHF-tau as fibrils or in its abnormal mobility on SDS gels. A l u m i n u m treatment of 3T3 fibroblasts transfected with h u m a n tau resulted in toxicity, but did not change tau expression levels or induce tau aggregation. In conclusion, aluminum appears to induce isolated tau protein to aggregate in a phosphate-independent way, without the formation of fibrils. This effect was not observed when tau-transfected cells were treated with toxic doses of aluminum.
INTRODUCTION Aluminum is a neurotoxin that induces the formation of neurofibrillary tangles in vivo and in cultured cells t. These tangles appear to contain aggregates or bundles of hyperphosphorylated neurofilaments 3'35. Purified neurofilaments aggregate in the presence of aluminum and show reduced electrophoretic mobility on SDS-PAGE when incubated with aluminum 22'27'33. In addition, neurofilament phosphopeptides adopt /3pleated sheet conformations in the presence of aluminum 15. The neurofibrillary tangles of Alzheimer's disease are comprised of paired helical filaments (PHFs), whose morphology differs from the straight fibrils present in aluminum-induced tangles 1. Alzheimer PHFs are comprised of tau protein, rather than neurofilaments 37. Nevertheless, many of the characteristics of aluminumtreated neurofilaments are similar to those of tau protein in Alzheimer's disease. For instance, PHFs contain hyperphosphorylated tau in an aggregated (fibril-
* Corresponding author. Fax: (1) (302) 886-2766.
lar) state 21 and show extensive /3-pleated sheet structure 17. Tau extracted from PHFs also shows decreased electrophoretic mobility on SDS-PAGE 21. Aluminum has been detected in Alzheimer tangles ~4. The aluminum chelator, deferoxamine, has been reported to have therapeutic effects on Alzheimer dementia and to release aluminum from body stores in Alzheimer patients t 1. Although it is not known whether the deferoxamine treatment releases aluminum from tangles, these results imply that aluminum may play a role in Alzheimer dementia. PHFs are likely to be a major contributor to Alzheimer dementia 4'6. Therefore any effect of aluminum to induce or stabilize PHFs would be expected to worsen dementia. The significance of the presence of aluminum in Alzheimer tangles is controversial. Its presence suggests the possibility that aluminum may play a role in initiating the formation of PHFs 28, analogous to the role proposed for aluminum in Alzheimer plaque formation 5. Aluminum is known to bind to phosphoryl groups on proteins 2. Therefore a potential mechanism for this initiation of P H F formation by aluminum is the binding of aluminum to the phosphate groups on tau, inducing it to polymerize. The feasibility of this possi-
78 bility is suggested by the observation that phosphorylated neurofilaments are induced to aggregate by aluminum and adopt a /3-pleated sheet conformation ~s'22. However, another possibility is that aluminum binds to PHF components, before or after their assembly into fibrils, without inducing their formation or stabilizing them. We performed a series of experiments to determine whether aluminum may play a role in the assembly of tau into PHFs. In particular we investigated whether aluminum may be involved in causing tau to aggregate or form fibrils or to stabilize the P H F structure. We treated non-phosphorylated recombinant human tau, phosphorylated bovine tau, hyperphosphorylated PHFtau and tau transfected fibroblasts with aluminum under conditions which induce neurofilament aggregation. PHF-tau and the aluminum-treated tau samples were incubated with deferoxamine to evaluate the reversible nature of the aggregates. The aggregation and fibril assembly properties of these samples were evaluated using electron microscopy and SDS-PAGE. MATERIALS
AND METHODS
Tau proteins Bovine tau was isolated as described 29. The shortest isoform of human tau (7352), and a tau fragment (z264), which encompasses amino acids 264-386 of the longest human tau isoform, were expressed in E. coli and purified as previously described 3°'31. SDSsoluble PHFs, referred to as A68 PHFs 21, were prepared from Alzheimer cortical tissue by a modification 9 of the procedure of Ksiezak-Reding et al. 19. Cell cultures 3T3 fibroblasts were stably transfected with cDNA for the shortest isoform of human tau (~-352) as previously described 23. Cells were plated in 100 mm Nunc culture dishes or onto glass coverslips and grown in D M E M (Gibco) supplemented with 10% fetal bovine serum. From 1 to 10 mM aluminum lactate was added to the medium for 1-3 days. Viability of coverslip cultures was assessed using the Live/Dead viability/cytotoxicity assay kit (Molecular Probes, Inc.). Cell extracts were prepared as previously described 34. Protein content was measured by the method of Lowry et al. 24 after TCA precipitation of extracted protein. Tau content was determined by ELISA7 using antibody tau-14, diluted 1:6,000 and purified recombinant human 7352 at 0.3-300 nM to generate the standard curve. Monoclonal antibody z-14, which recognizes amino acids 141-178 of ~-441TM was generously provided by Dr. Virginia Lee. Other methods and reagents SDS-gel electrophoresis was performed on isolated proteins and on cell extracts and analyzed by Western blot analysis 34 and PROBLUE staining of proteins (Enprotech). Electron microscopy of protein samples that were negatively stained with 2% uranyl acetate and PHF samples negatively stained with 1% lithium phosphotungstic acid was performed as previously described 8. Deferoxamine (Desferal, CIBA Pharmaceutical Company) was hydrated in water prior to use. Aluminum perchlorate, aluminum lactate, aluminum sulfate, chromium chloride, ferric chloride and magnesium chloride were of reagent grade and were prepared as stock solutions immediately prior to use.
RESULTS
Bovine tau was incubated with increasing concentrations of aluminum perchlorate and then electrophoresed using reducing SDS-PAGE. As shown in Fig. 1, bovine tau protein migrated as three to five bands. Samples of bovine tau that were incubated with aluminum showed decreased electrophoretic mobility. The typical tau bands disappeared and were replaced with a smear of protein covering the 50-100 kDa range. In some samples a faint band at 130 kDa was also observed. This pattern suggests that aluminum induced tau conformations that were stable to SDS and boiling and appeared as slowly migrating species on reducing SDS gels. This effect was dependent on both the aluminum and tau concentrations. With 10 /xM bovine tau, the lowest effective dose of aluminum that caused a mobility shift was 200 /~M. At higher aluminum concentrations, further decreases in tau mobility were seen. At aluminum concentrations above 800 /~M, tau did not enter the gel matrix, presumably due to the formation of SDS-insoluble aggregrates. These effects are similar to those reported with aluminum and neurofilaments 22. Incubating the aluminum-treated tau samples with deferoxamine reversed the effect of aluminum on tau mobility, causing tau to migrate similar to the untreated control sample (Fig. 1B). Presum-
A
B 1 2
3
4
5 6
7
1
2
3
4
5
6
66-4.1--
22-14-Fig. 1. Effect of aluminum on the electrophoretic mobility of bovine tau. A: bovine tau (10/~M) was incubated alone (lane 2) or with 0.2 mM (3), 0.4 mM (4), 0.6 mM (5), 0.8 mM (6), or 1 mM aluminum perchlorate (7) at pH 6.7 for 30 min at 37°C. The samples were heated in sample buffer (62.5 mM Tris, pH 6.9, 1% SDS, 10% glycerol, 20 mM DTT, 0.005% Bromphenol blue), electrophoresed on a 10-20% gradient polyacrylamide gel and the gel stained with PRO-BLUE (Enprotech). Lane 1 contains molecular weight standards. B: bovine tau (10/xM) was incubated alone (lanes 1, 4) or with 0.4 mM (lanes 2, 5) or 1.0 mM aluminum perchlorate (lanes 3, 6) at pH 6.7 for 30 min at 37°C. Some samples were then incubated with 1 mM deferoxamine for 30 min (lanes 4-6). Samples were then heated in sample buffer and analyzed as described for A.
79
B
A
1
2
34
66
7
1
8
2 3 4 6 6
97-U-4S-66--
31--
43--
22-14--
31--
6--
14
g
Fig. 2. Effect of aluminum on the electrophoretic mobility of recombinant tau and a recombinant tau fragment. A: r352 (10 p.M) was incubated alone (lane 2) or with 0.2 mM (3), 0.4 mM (4), 0.6 mM (5), 0.8 mM (6), 1 mM (7) or 3 mM aluminum perchlorate (8) at pH 6.7 for 30 min at 37°C. Samples were analyzed as described in Fig. 1A. B: r264 (10 p~M) was incubated alone (lane 2) or with 0.2 mM (3), 0.5 mM (4), 1 mM (5) or 5 mM aluminum perchlorate (6) at pH 6.7 for 30 min at 37°C. Samples were heated in sample buffer, electrophoresed on a 10-20% gradient gel using Tris/Tricine buffer system (Enprotech), and stained with PRO-BLUE.
ably, deferoxamine was able to remove the aluminum which, when bound to tau, caused the change in tau mobility. Recombinant human tau (r352) was incubated with increasing concentrations of aluminum perchlorate and analyzed by SDS gel electrophoresis. As with the bovine tau samples, the aluminum-treated recombinant human tau samples ran as a smear of protein covering the 50-80 kDa range (Fig. 2A). A tau band at approximately 130 kDa was also observed. At aluminum concentrations greater than 1 mM tau formed aggregates that did not enter the gel matrix. Since recombinant tau is not phosphorylated 32, this result suggests that aluminum binds to the amino acid side chains of tau rather than to phosphoryl groups to elicit its effect on tau mobility. This effect was not dependent on the aluminum salt, as aluminum lactate and aluminum sulfate were also capable of inducing aggregation of bovine and recombinant tau. Ferric chloride did not cause a shift in tau mobility, although at 10 mM it did appear to create tau aggregates that did not enter the gel (data not shown). Ten mM magnesium chloride had no effect on tau mobility. Experiments were performed to see if aluminum could induce the aggregation of r264, a recombinant tau fragment. This fragment contains the microtubule binding domains of tau and encompasses the region of tau that forms the protease-resistent core of SDS-insoluble PHFs 16'29. Concentrations of aluminum that caused the shift in electrophoretic mobility of bovine
tau and recombinant tau to the 50-80 kDa range had no effect on the electrophoretic mobility of the tau fragment (Fig. 2B). However, at aluminum concentrations of 1 mM or greater the tau fragment failed to enter the gel matrix, presumably due to its aggregation into SDS-insoluble structures. Thus, the small shift in gel migration to 50-80 kDa was specific to full length taus and required protein sequence outside of the microtubule binding domain. The formation of large aggregates that do not electrophorese was a property of both full length taus and the tau fragment. A68 PHFs were isolated and experiments performed to see whether aluminum or deferoxamine affects the aggregation state of these fibrils. A68 PHFs are soluble in SDS, and the proteins that comprise these filaments can be separated by SDS gel electrophoresis 2~. The major component of PHFs is hyperphosphorylated tau protein 2t. PHF-tau shows reduced mobility on SDS gels, probably as a consequence of its abnormal phosphorylation state ~3. To determine whether aluminum may also contribute to the reduced mobility of PHF-tau, samples of A68 PHFs were treated with deferoxamine and then analyzed by SDS-PAGE. No increased mobility of PHF-tau was observed using 1-20 times the concentration of deferoxamine that was effective in reversing the aluminum induced mobility change with bovine and recombinant tau (Fig. 3). This suggests that either aluminum is not responsible for the reduced mobility of PHF-tau or aluminum (if it is bound to PHF-tau) has a much higher affinity for PHF-tau than recombinant or bovine tau, preventing its removal by deferoxamine. Interestingly, PHF-tau was similar to the other tau preparations in terms of its susceptibility to forming aggregates with aluminum that were stable to SDS treatment (Fig. 3). This effect was observed with both isolated A68 PHFs and SDS-solubilized PHF-tau. Samples of bovine tau, r352 and r264 that showed aluminum-induced mobility changes were examined by negative stain electron microscopy. No fibrillar structures were observed in these aluminum-treated samples (Fig. 4). Both aluminum and deferoxamine-treated A68 PHFs were also examined by electron microscopy. Neither treatment altered the number of fibrils present in the A68 P H F samples, or the morphology of the PHFs (data not shown). It is possible that a nucleating agent is required for aluminum-tau polymerization. Such an agent would not be present in studies using purified tau protein, but may be present in intact cells. Therefore, experiments were performed to see whether aluminum could induce the formation of tau fibrils in tau-transfected 3T3 fibroblasts. The morphology of tau-transfected 3T3 cells
80 is altered, with cells expressing ~-352 appearing more flattened than non-transfected fibroblasts z3. Aluminum lactate treatment produced a dose-dependent toxic effect on the transfected cells. The number of cells that stayed attached to the plate decreased as a function of aluminum concentration (Fig. 5), and the amount of protein extracted per culture also decreased (Table I). The morphology of the cells that remained attached also changed upon aluminum treatment (Fig. 5), with thin processes becoming prominent. These processes appear similar to those that are seen on 3T3 fibroblasts expressing other tau isoforms 23. The cells exhibiting these processes were still viable, as assessed by the L i v e / D e a d assay (Fig. 5). Toxicity was apparent after two days of treatment of low density cultures with 4 mM aluminum, while cells cultured at a higher density appeared slightly less sensitive to the treatment. Although aluminum treatment resulted in a decrease in the number of attached, viable cells and the total protein content of the cell extract, it did not cause a decrease in the level of tau expression based on percentage of total protein (Table I). In fact, at the higher aluminum doses tau expression accounted for a
1
2
3
4
97~ 68-3--
14-Fig. 3. Western blot showing the effects of aluminum and deferoxamine on the mobility of PHF-tau. Lane 1 contains bovine tau. Aliquots of A68 PHFs were incubated either alone (lane 2) or with 10 mM desferoxamine (lane 3) or 1 mM aluminum perchlorate (lane 4) under the conditions described in Fig. 1. The samples were electrophoresed then transferred to nitrocellulose and probed with antibody ~-14. PHF-tau was used at a concentration similar to bovine tau and recombinant tau, based on relative immunostaining intensities.
TABLE 1
Effect of aluminum lactate treatment of "r-transfected 3T3 cells on protein expression Cells were cultured at two densities and treated with aluminum for 3 days. Cell extracts were analyzed for total protein content and tau protein content as described in Materials and Methods.
Cell density
Aluminum concerttration (mM)
Total protein Ozg / lO0 mm dish)
Tau protein (% of total protein)
(n)
1×
0 2 4 6 8
130 121 79 42 18
3.9±1.6 2.8±0.8 1.9±0.5 2.1±0.5 4.8±1.4
6 6 6 5 4
0.5 ×
0 2 4 6 8
105 99 42 24 19
2.7±0.6 3.5±0.9 1.5±0.7 3.7±0.8 4.4±1.1
6 6 3 4 2
larger portion of the total protein. Aliquots of each extract, which contained equivalent protein contents and tau contents (based on ELISA), were applied to gels and electrophoresed to assess the state of aggregation of the expressed tau protein. No change in the electrophoretic mobility of tau was detected in the aluminum-treated cell extracts, compared to the nontreated cell extracts, when extract samples containing either 5 p.g (Fig. 6) or 10/xg protein (data not shown) were electrophoresed. The distribution of tau between the two bands, which represents different phosphorylation states of 734, was not affected by aluminum treatment. DISCUSSION It has been suggested that aluminum binds to the negatively charged phosphate groups on phosphoproteins such as neurofilaments, and thereby induces crosslinking and aggregation12'2% In this study we evaluated several different tau proteins which are phosphorylated to different extents, to determine whether the phosphorylation state altered the sensitivity of tau to form aggregates or fibrils. Recombinant human tau isoform ~-352 and tau fragment r264 are isolated as nonphosphorylated proteins 31 while bovine tau contains at least three phosphate groups 29. Hyperphosphorylated PHF-tau isolated from A68 PHFs contains up to eight phosphate groups 2° and human tau expressed in 3T3 fibroblasts migrates as two bands on gels, reflecting heterogenous phosphorylation states 3a. Aluminum caused the isolated tau proteins to aggregate, as defined by a decrease in migration on SDSP A G E and a loss of protein entering the gel matrix.
81 These effects are analogous to the aluminum effects on neurofilaments and occur with similar aluminum concentrations 22'27. The aluminum-induced tau aggregates were not fibrillar or PHF-like in morphology, but rather appeared as amorphous aggregates. Nonphosphorylated r352 and bovine tau were equally sensitive to aluminum, indicating that phosphorylation of tau at physiological sites does not enhance the aluminum effect. The recombinant fragment r264 formed aggregates in the presence of aluminum that failed to enter the gel matrix, but did not show the decreased mobility seen with intact tau. Presumably, regions outside of the microtubule binding domains are required to achieve this conformation. It cannot be discerned from these data whether this conformation reflects an intermolecular aggregation of tau monomers or an intramolecular conformation that migrates anomalously on SDS gels. Aluminum was toxic to fibroblasts that expressed tau. However tau did not aggregate in these cells, eliminating it as a mechanism for the observed toxicity.
In contrast, neurofilaments do aggregate in aluminumtreated cells =. The transfected fibroblasts treated with the highest doses of aluminum showed an increase in tau content. The increased tau content was probably not responsible for the toxicity, as cell death was observed with concentrations of aluminum below that which caused an increase in tau content. Furthermore, an aluminum-induced increase in tau immunoreactivity has been observed in human neuroblastoma ceils, without an effect on cell viability 26. Other mechanisms of aluminum toxicity that may have been responsible for the toxicity we observed include interference of aluminum with D N A function 1° or with magnesium-mediated biological processes, as aluminum and magnesium have similar ionic radii and can compete in various biological systems 35. Aluminum causes both tau and neurofilament monomers in soluble form to aggregate into non-fibrillar material. Aluminum also appears to induce neurofilaments in fibril form to bundle, forming tangles z2'36,
Fig. 4. Electron micrographs of PHFs (A) and aluminum treated tau preparations, r352 (B), r264 (C) and bovine tau (D) were incubated with 1 mM aluminum perchlorate for 30 rain at 37°C and then placed on carbon-formvar coated grids, stained with 1% lithium phosphotungstic acid and examined by electron microscopy.Magnification, 74,250 ×.
82
~
i¸
~ . . . .
e
~ ~
~
-
¸
Fig. 5. Phase contrast micrographs of 3T3 cells transfected with h u m a n r352 D N A and cultured in the absence (A), or presence of aluminum lactate at 2 rnM (B), 4 m M (C), 6 m M (D), or 8 m M (E). Arrowheads indicate dead cells as determined by L i v e / D e a d assay of the same field. No cells were dead in B, one cell was dead in A, C, and D; most of the cells were dead in E.
83 1
2
3
4
5
6
7
8
9
10
11
SSm 43--
2w
Fig. 6. Western blot of extracts of the cell cultures in Fig. 5 are shown. Aliquots of extracts, containing 5 /zg of protein, were electrophoresed, transferred to nitrocellulose and probed with antibody tau-14 for tau immunoreactivity. Lane 1 contains purified human ~-352 and ~-441 expressed in E. coll. Lanes 2-6 contain extracts of cells cultured at an initial plating density that was twice that of the cultures used for the extracts in lanes 7-11. Cultures were untreated (lanes 2, 7) or treated with aluminum at 2 mM (lanes 3, 8), 4 mM (lanes 4, 9), 6 mM (5, 10), or 8 mM (6, 11).
but does not induce similar tau-containing tangles to form under the conditions used in this study. Since neurofilament monomers normally assemble into fibrils, the effect of aluminum in vivo is to cause aggregation of the assembled fibrils of neurofilaments into large aggregates or tangles. However since tau does not normally form fibrils, aluminum treatment does not produce aggregates or tangles of tau fibrils nor does it act as a nucleating agent under the conditions tested to induce tau to form fibrils. The lack of effect of deferoxamine treatment on P H F morphology and electrophoretic mobility of PHF-tau suggests that aluminum is not required to maintain the P H F structure, nor is it responsible for the abnormal mobility of PHF-tau. Acknowledgments. The authors wish to acknowledge the technical assistance of Hope Koenigbauer, William Brunner and Irene Evangelista Sobel. The authors thank Dr. Claude Wischik and the Cambridge Brain Bank for providing Alzheimer cortical brain tissue and Dr. Michel Goedert for supplying the tau clones.
3 Bizzi, A. and Gambetti, P., Phosphorylation of neurofilaments is altered in aluminum intoxication, Acta Neuropathol., 71 (1986) 154-158. 4 Braak, H. and Braak, E., Neuropathological stageing of Alzheimer-related changes, Acta Neuropathol., 82 (1991) 239-259. 5 Candy, J.M., Oakley, A.E., Klinowski, J., Carpenter, T.A., Perry, R.H., Atack, J.R., Perry, E.K., Blessed, G., Fairbairn, A. and Edwardson, J.A., Aluminosilicates and senile plaque formation in Alzheimer's disease, Lancet, 2 (1986) 354-357. 6 Caputo, C.B. and Salama, A.I., The amyloid proteins of Alzheimer's disease as potential targets for drug therapy, Neurobiol. Aging, 10 (1989) 451-461. 7 Caputo, C.B., Wischik, C., Novak, M., Scott, C.W, Brunner, W.F., Montejo de Garcini, E., Lo, M.M.S., Norris, T.E. and Salama, A.I., Immunological characterization of the region of tau protein that is bound to Alzheimer paired helical filaments. Neurobiol. Aging, 13 (1992) 267-274. 8 Caputo, C.B., Sygowski, L.A., Scott, C.W and Sobel, I.R.E., Role of tau in the polymerization of peptides from /3-amyloid precursor protein, Brain Res., 597 (1992) 227-232. 9 Caputo, C.B., Hayward, C., Brunner, W.F., Scott, C.W and Sygowski, L.A., Association of the C-terminus of fl-amyloid precursor protein with a tau protein fragment in Alzheimer PHF material, Biochem. Biophys. Res. Commun., 185 (1992) 1043-1040. 10 Crapper McLachlan, D.R., Lukiw, W.J. and Kruck, T.P.A., New evidence for an active role of aluminum in Alzheimer's disease, Can. J. Neurol. Sci., 16 (1989) 490-497. 11 Crapper McLachlan, D.R., Dalton, A.J., Kruck, T.P.A., Bell, M.Y., Smith, W.L., Kalow, W. and Andrews, D.F., Intramuscular deferoxamine in patients with Alzheimer's disease, Lancet, 337 (1991) 1304-1308. 12 Ganrot, P.-O., Biochemistry and metabolism of AICI 3 and similar ions: a review, Environ. Health Perspect., 65 (1986) 363-441. 13 Goedert, M., Spillantini, M.G., Cairns, N.J. and Crowther, R.A., Tau proteins of Alzheimer paired helical filaments: abnormal phosphorylation of all six brain isoforms, Neuron, 8 (1992) 159168. 14 Good, P.F., Perl, D.P., Bierer, L.M. and Schmeidler, J., Selective accumulation of aluminum and iron in the neurofibrillary tangles of Alzheimer's disease: a laser microprobe (LAMMA) study, Ann. Neurol., 31 (1992) 286-292. 15 Hollosi, M., Urge, L., Perczel, A., Kajtar, J., Teplan, I., Otvos, L. and Fasman, G.D., Metal ion-induced conformational changes of phosphorylated fragments of human neurofilament (NF-M) protein, J. Mol. Biol., 223 (1992) 673-682. 16 Jakes, R., Novak, M., Davison, M. and Wischik, C.M., Identification of 3- and 4-repeat tau isoforms within the PHF in Alzheimer's disease, EMBO J., 10 (1991) 2725-2729. 17 K.irschner, D.A., Abraham, C. and Selkoe, D.J., X-ray diffraction from intraneuronal paired helical filaments and extraneuronal amyloid fibers in Alzheimer disease indicates cross-/3 conformation, Proc. Natl. Acad. Sci. USA, 83 (1986) 503-507. 18 Kosik, K.S., Orecchio, L.D., Binder, L., Trojanowski, J.Q., Lee, V.M.-Y. and Lee, G., Epitopes that span the tau molecule are shared with paired helical filaments, Neuron, 1 (1988) 817-825. 19 Ksiezak-Reding, H., Binder, L.I. and Yen, S.-H., Alzheimer disease proteins (A68) share epitopes with tau but show distinct biochemical properties, J. Neurosci. Res., 25 (1990) 420-430. 20 Ksiezak-Reding, H., Liu, W.-K. and Yen, S.-H., Phosphate analysis and dephosphorylation of modified tau associated with paired helical filaments, Brain Res., 597 (1992) 209-219. 21 Lee, V.M.-Y., Balin, B.J., Otvos, L. and Trojanowski, J.Q., A68 a major subunit of paired helical filaments and derivatized forms of normal tau, Science, 25 (1991) 675-678. 22 Leterrier, J.F., Langui, D., Probst, A. and Ulrich, J., A molecular mechanism for the induction of neurofilament bundling by aluminum ions, J. Neurochem., 58 (1992) 2060-2070. 23 Lo, M.M.S., Fieles, A.W., Norris, T.E., Dargis, P.G., Caputo, C.B., Scott, C.W, Lee, V.M.-Y. and Goedert, M., Human tau isoforms confer distinct morphological and functional properties to stably transfected fibroblasts, Mol. Brain Res., 20 (1993) in Dress. -
REFERENCES 1 Bertholf, R.L., Aluminum and Alzheimer's disease: Perspectives for a cytoskeletal mechanism, Crit. Rev. Clin. Lab. Sci., 25 (1987) 195-210. 2 Birchali, J.D. and Chappell, J.S., The chemistry of aluminum and silicon in relation to Alzheimer's disease, Clin. Chem., 34 (1988) 265-267.
-
84 24 Lowry, O.H., Rosebrough, N.J., Farr, A,L., and Randall, R.J., Protein measurement with the Folin phenol reagent, J. Biol. Chem., 193 (1951) 265-27. 25 Macdonald, T.L. and Martin, F.B., Aluminum ion in biological systems, Trends Biol. Sci., 13 (1988) 15-19. 26 Mesco, E.R., Kachen, C. and Timiras, P.S., Effects of aluminum on tau proteins in human neuroblastoma cells, Mol. Chem. Neuropathol., 14 (1991) 199-212. 27 Nixon, R.A., Clarke, J.F., Logvinenko, K.B., Tan, M.K.H., Hoult, M. and Grynspan, F., Aluminum inhibits calpain-mediated proteolysis and induces human neurofilament proteins to form protease-resistent high molecular weight complexes, J. Neurochem., 55 (1990) 1950-1959. 28 Perl, D.P. and Pendlebury, W.W., Aluminum neurotoxicity - potential role in the pathogenesis of neurofibrillary tangle formation, Can. J. Neurol. Sci., 13 (1986) 441-445. 29 Poulter, L., Barratt, D., Scott, C.W and Caputo, C.B., Locations and immunoreactivities of phophorylation sites on bovine and porcine tau proteins and a PHF-tau fragment, J. Biol. Chem., 268 (1993) 9636-9644. 30 Scott, C.W, Blowers, D.P., Barth, P.T., Lo, M.M.S., Salama, A.I. and Caputo, C.B., Differences in the abilities of human tau isoforms to promote microtubule assembly, J. Neurosci. Res., 30 (1991) 154-162. 31 Scott, C.W, Klika, A.B., Lo, M.M.S., Norris, T.E. and Caputo, C.B., Tau protein induces bundling of microtubules in vitro:
32
33
34
35
36
37
comparison of different tau isoforms and a tau protein fragment, J. Neurosci. Res., 33 (1992) 19-29. Scott, C.W, Spreen, R.C., Herman, J.L., Chow, F.P., Davison, M.D., Young, J. and Caputo, C.B., Phosphorylation of recombinant tau by cAMP-dependent protein kinase, J. Biol. Chem., 268 (1993) 1166-1173. Shea, T.B., Beerman, M.L. and Nixon, R.A., Aluminum alters the electrophoretic properties of neurofilament proteins: role of phosphorylation state, J. Neuroehem., 58 (1992) 542-547. Sygowski, L.A., Fieles, A.W., Lo, M.M.S., Scott, C.W and Caputo, C.B., Phosphorylation of tau protein in tau-transfected 3T3 cells, Mol. Brain Res., 20 (1993) in press. Troncosco, J.C., Sternberger, N.H., Sternberger, L.A., Hoffman, P.N. and Price, D.L., Immunocytochemical studies of neurofilament antigens in the neurofibrillary pathology induced by aluminum, Brain Res., 364 (1986) 295-300. Troncoso, J.C., March, J.L., Haner, M. and Aebi, U., Effect of aluminum and other multivalent cations on neurofilaments in vitro: An electron microscopic study, J. Struct. Biol., 103 (1990) 2-12. Wischik, C.M., Novak, M., Thogersen, H.C., Edwards, P.C., Runswich, M.F., Jakes, R., Milstein, C., Roth, M. and Klug, A., Isolation of a fragment of tau derived from the core of the paired helical filament of Alzheimer disease, Proc. Natl. Acad. Sci. USA 85 (1988) 4506-4510.