Amyloid-like aggregates of neuronal tau induced by formaldehyde promote apoptosis of neuronal cells

Background The microtubule associated protein tau is the principle component of neurofibrillar tangles, which are a characteristic marker in the pathology of Alzheimer's disease; similar lesions are also observed after chronic alcohol abuse. Formaldehyde is a common environmental contaminant and also a metabolite of methanol. Although many studies have been done on methanol and formaldehyde intoxication, none of these address the contribution of protein misfolding to the pathological mechanism, in particular the effect of formaldehyde on protein conformation and polymerization. Results We found that unlike the typical globular protein BSA, the natively-unfolded structure of human neuronal tau was induced to misfold and aggregate in the presence of ~0.01% formaldehyde, leading to formation of amyloid-like deposits that appeared as densely staining granules by electron microscopy and atomic force microscopy, and bound the amyloid-specific dyes thioflavin T and Congo Red. The amyloid-like aggregates of tau were found to induce apoptosis in the neurotypic cell line SH-SY5Y and in rat hippocampal cells, as observed by Hoechst 33258 staining, assay of caspase-3 activity, and flow cytometry using Annexin V and Propidium Iodide staining. Further experiments showed that Congo Red specifically attenuated the caspase-3 activity induced by amyloid-like deposits of tau. Conclusion The results suggest that low concentrations of formaldehyde can induce human tau protein to form neurotoxic aggregates, which could play a role in the induction of tauopathies.


Background
Although many studies have been done on methanol and formaldehyde intoxication [1,2], none of these address the contribution of protein misfolding to the pathological mechanism, in particular the effect of formaldehyde on protein conformation and polymerization. Damage of neuronal cells caused by misfolded protein aggregates is a subject of intense research interest. It has become increasingly clear that many neurodegenerative diseases are related to aggregation and deposition of misfolded proteins, such as tau [3][4][5], beta amyloid [6][7][8][9], alpha-synuclein [10,11] and polyglutamine aggregates [12,13]. The abnormal deposition of misfolded protein causes the malfunction of a distinctive set of neurons [14]. Alzheimer's disease and some other dementias are related to pathological deposition of proteins. Tau is a microtubule-associated protein, which is the main constituent of paired helical filaments (PHFs) present in neurofibrillary tangles [4,5]. In neurodegeneration, tau protein accumulates in lesions composed of fibrillar aggregates displaying the cross β-sheet diffraction pattern of "amyloid" [15]. Interestingly, neurofibrillary tangles have been found in brains of chronic alcoholics possessing neuropathological signs of thiamine-deficiency, suggesting that tau misfolding may be involved in the alcohol-induced pathological pathway [16][17][18].
Methanol ingestion is an important public health concern because of the selective actions of its toxic metabolites, formaldehyde and formic acid, on the retina, optic nerve and central nervous system [1]. Severe and even fatal illness has been reported after illicit consumption of "industrial methylated spirits" [2]. Methanol is oxidized by alcohol dehydrogenase to produce formaldehyde, which is further oxidized to formic acid by formaldehyde dehydrogenase. Metabolism of methanol to formaldehyde via peroxisomal enzymes has been demonstrated in rat retina in vitro [19], and the presence of cytoplasmic aldehyde dehydrogenase activity has been demonstrated in several regions of the rat and mouse eye, including the retina [20,21]. Susceptibility to methanol toxicity is dependent upon the relative rate of formate clearance. However, methanol toxicosis induces progressive damage to the central nervous system. It is hard to explain this chronic damage by local accumulation of formic acid alone.
Formaldehyde is a common environmental contaminant found in paint, clothes, medicinal and industrial products, and is a component of diesel and gasoline exhaust [22,23]. Recently, Sarsilmaz and colleagues have reported that formaldehyde exposure may cause various morphological changes in the rat brain [24,25]. Neurotoxic effects have been also confirmed by acute and subacute formaldehyde exposure in mice [26]. Pitten et al. have classified formaldehyde as "probably neurotoxic" [27], because they found rats exposed to formaldehyde need more time and make more mistakes than the animals of the control group while going through a maze. As a crosslinking agent, formaldehyde readily reacts with thiol and amino groups [28], causing polymerization of proteins. In semicarbazide-sensitive amine oxidase (SSAO)-mediate pathogenesis of Alzheimer's disease, formaldehyde interacts with β-amyloid and produces irreversibly cross-linked neurotoxic amyloid-like complexes [29][30][31]. Therefore, the potential effect of formaldehyde on protein misfolding may be significant, even if formaldehyde remains in the human body for only a short time.
Here, we examine the role of formaldehyde in induction of protein misfolding. In particular, we investigate the effect of formaldehyde on the aggregation of human neuronal tau in vitro and the toxicity of tau aggregates in mammalian neuronal cells. The results imply that low concentrations of formaldehyde are sufficient to induce formation of amyloid-like tau aggregates, which in turn induce apoptosis of both human neuroblastoma cells (SH-SY5Y) and rat hippocampal cells.

Formaldehyde at low concentrations induces tau to form amyloid-like aggregates
In order to investigate the potential of formaldehyde to induce protein misfolding leading to neurodegenerative disease, we studied the effect of low concentrations of formaldehyde on the biochemical and biophysical properties of human neuronal tau. Incubation with increasing concentrations of formaldehyde from 0.01-0.5% was observed to result in formation of increasing amounts of SDS-insoluble aggregates of tau, as detected by SDS-PAGE ( Fig. 1A; Fig. 1C, curve 1) or light scattering (Fig. 1C, curve 2), similar to our previous study [32,33]. In contrast, when BSA was incubated with formaldehyde under the same range of conditions, no significant degree of aggregation could be detected by SDS-PAGE ( Fig. 1B; Fig. 1D, curve 1) or by light scattering (Fig. 1D, curve 2). This suggests that tau is more susceptible to the effects of formaldehyde than typical globular proteins.
We then compared the time course of the aggregation reaction in the presence (Fig. 1E, curve 1) and absence (Fig. 1E, curve 2) of 0.1% formaldehyde by monitoring the degree of light scattering. In both cases a sigmoidal curve was observed. However, the presence of formaldehyde both reduced the lag time for aggregation and resulted in a greater degree of light scattering (Fig. 1F). The presence of formic acid had no detectible effect on tau aggregation (data not shown). This further suggests that formaldehyde promotes aggregation of tau. In contrast, incubation of BSA under the same conditions showed little aggregation in the absence (Fig. 1E, curve 3) or presence (Fig. 1E, curve 4) of formaldehyde.
OPT is commonly used as a fluorescent probe to detect both αand ε-amino groups of a protein [34]. As shown in Fig. 2A, the fluorescent intensity at 455 nm decreased as the formaldehyde concentration increased. In 1% formaldehyde solution, the fluorescence could hardly be detected under the same conditions, indicating that formaldehyde competed with OPT in the reaction with the amino groups. The time course of the fluorescence changes in the presence and absence of 0.005% formaldehyde monitored using OPT showed a marked difference (Fig. 2B). The first order rate of the fluorescence change in Effect of formaldehyde on tau aggregation at different concentrations  Light Scattering (Arbitrary Units) C E D the absence of formaldehyde (~1.03 × 10 -3 s -1 ) was greater than that in the presence of formaldehyde (~3.66 × 10 -4 s -1 ). This demonstrates that formaldehyde reacts with protein tau and blocks amino groups.
The time course of tau aggregation in the presence and absence of formaldehyde was also monitored using the fluorescent dye ThT, which was considered to be highly specific for amyloid-like structure [35]. Under the conditions used, self-aggregated tau showed only minor change in the ThT fluorescence (curve 2, Fig. 3A and 3B). Acetaldehyde-treated tau likewise did not show any significant change in the ThT spectrum (curve 3, Fig. 3A), compared with a protein-free control (curve 4, Fig. 3A). Interestingly, however, formaldehyde-induced aggregation of tau showed a significant change in the intensity of ThT fluorescence (curve 1, Fig. 3A), and the sigmoidal time course of aggregation detected by ThT binding (curve 1, Fig. 3B) was similar to that observed by light scattering (curve 1, Fig. 1E). Formaldehyde-treated tau was also found to bind Congo Red, causing a red shift in the spectral maximum from 470 nm to 510 nm (curve 1, Fig. 3C), as observed for amyloidogenic tau peptides [36], compared with acetaldehyde-treated tau, native tau or Congo Red alone as controls (curves 2, 3 and 4, Fig. 3C). Negatively-stained electron micrograph images of the formaldehyde-induced aggregates showed dense granules (Fig. 4A), whereas selfaggregated tau as control showed PHF-like structures by transmission electron microscopy ( Fig. 4B), similar to the results reported previously [5]. By AFM, protein tau treated with 0.05% formaldehyde also showed globular particles on the mica surface. The horizontal diameter of the tau particles was 18.65 ± 2.66 nm (mean ± SD, Fig.  4C) about twice the size of native tau (9.94 ± 1.96 nm, Fig.  4D). However, fibril-like aggregates could not be observed in the formaldehyde-treated tau that was incubated for over a week.
Consequentially, we measured the activity of the different types of tau aggregates in tubulin assembly. While the spontaneously formed aggregates and those formed in the presence of acetaldehyde, maintained a relatively high level of residual activity, the formaldehyde-induced aggregates of tau were practically inactive in promoting tubulin assembly (Table 1). Together, these results indicate that the presence of formaldehyde promotes the formation of amyloid-like aggregates of tau, causing tau to become inactive in tubulin assembly.

Conformational changes of amyloid-like tau aggregates
The ability of tau aggregates to bind the dyes ThT and Congo Red suggests that the aggregated tau contains relatively more β-sheet structure compared with the natively unfolded protein. To prove this prediction, htau 40 was incubated in the presence or absence of formaldehyde and then examined by circular dichroism (CD) spectroscopy.
In the absence of formaldehyde, both native tau and acetaldehyde-treated tau showed CD spectra typical for an unfolded protein as reported [37,38], with a broad minimum of ellipticity centered at 205 nm (curves 2 and 3 respectively, Fig. 5A). The presence of formaldehyde led to a noticeable change in the spectra so that the minimum became wider and was shifted toward higher wavelength, suggesting a substantial change from random coil to β structure consistent with the premolten globule folding state (curve 1, Fig. 5A), similar to previous reports of an increase in β conformation in tau aggregation [37][38][39].
These data further support that formaldehyde-induced tau aggregates contain partially folded structure that is enriched in β-sheet content in contrast to the initial conformation.
In addition to its secondary structure signature, the premolten globule state is characterized by a partially collapsed structure with a loosely packed hydrophobic core.
To confirm that the aggregates had partially folded character, htau40 was incubated with formaldehyde and then examined for the ability to bind ANS, a fluorescent probe of surface-exposed hydrophobic patches [40]. ANS in buffer alone (not shown) fluoresced weakly at the optimum wavelength (Ex: 350 nm; Em: 480 nm) [39], which was unaffected by addition of increasing concentrations of formaldehyde in the absence of tau (curve 4, Fig. 5B).
In contrast, ANS in the presence of formaldehyde-aggregated tau fluoresced brightly (curve 1, Fig. 5B). A slight increase in fluorescence intensity was detected for selfaggregated tau (curve 2, Fig. 5B), indicating some PHFlike structures could form as described previously [5,32]. Note that no marked increase in fluorescence intensity was detected for acetaldehyde-treated tau (curve 3, Fig.  5B). This suggests that the enhanced fluorescent signal observed came from the binding of ANS to altered conformations of tau induced by formaldehyde.

Amyloid-like tau promotes apoptosis of neuronal cells
In order to investigate whether amyloid-like tau has an effect on neurons, amyloid-like tau was added to neurotypic SH-SY5Y cells after the residual formaldehyde had been removed completely from the protein-deposits by ultrafiltration (see Methods). Figure 6 illustrates that amyloid-like tau induces marked axonal atrophy and finally the cells shrink into a spherical shape in the neuroblastoma culture. To tell whether this deposit-induced cell death represented apoptosis, we examined two features of apoptosis: the morphology (nuclear condensation and fragmentation) and the biochemical (caspase activity) changes of the cells. As shown in Fig. 7, morphological evaluation of neuronal cultures using Hoechst 33258 staining with fluorescence microscopy revealed a significant increase in the number of cells showing nuclear con-Reaction of formaldehyde with amino groups of neuronal tau   densation and fragmentation after incubation with amyloid-like tau for 3 days (Fig. 7E). Under these conditions, 0.1% formaldehyde solution ultrafiltered with 100 mM phosphate buffer (pH 7.2) was used as a control. No marked characteristics of apoptosis could be detected by Hoechst 33258 (Fig. 7D) in the presence of the solution whose formaldehyde was removed by the ultrafiltration. Further, no effect on cell morphology could be detected in the presence of a similar dose of self-aggregated tau (Fig.  7B) or acetaldehyde-induced tau (Fig. 7C).
The effect of amyloid-like tau on cell viability was estimated using the MTT assay [41]. As shown in Fig. 8A, the percentage of viable cells in the culture decreased significantly over the course of 72 h after treatment with amyloid-like tau. In contrast, the presence of self-aggregated tau or acetaldehyde-treated tau had no effect on the number of viable cells under the same conditions. Over the same time period, the caspase-3 activity of cytosolic protein extracts from the SH-SY5Y cells treated with amyloid-like tau was observed to increase, while the caspase-3 activity for the control cultures remaining constant (Fig.  8B). The effects of Congo Red on the ability of formaldehyde-aggregated tau to induce apoptosis in the SH-SY5Y cell-culture were examined. The caspase-3 activity produced in response to treatment with formaldehyde-aggregated tau was reduced significantly by addition of Congo Red (Fig. 8C). Congo Red alone caused no detectable effect on cell viability (not shown). This then suggests that the ability of Congo Red to bind to amyloid-like tau confers a protective effect on neurons, as has been observed in the case of β-amyloid [42].
In order to further investigate the neurotoxicity of amyloid-like tau, we used flow cytometry analysis to examine the effect of formaldehyde-aggregated tau on rat hippocampal cells (Fig. 9). Cells were incubated with tau aggregates (after complete removal of formaldehyde traces, see Methods) and then treated with the dyes Annexin V (AV, an indicator of apoptosis) and propidium iodide (PI, an indicator of cell necrosis) [43]. A clear increase was observed in both the AV positive/PI negative (apoptotic) and AV-positive/PI positive (late apoptosis, early necrosis) populations of primary hippocampal cells during the time course of treatment with amyloid-like tau (Fig. 9B-D). We also investigated the efficiency of amyloid-tau produced by treatment with different concentrations of formaldehyde to induce apoptosis. Incubation for 72 h with amyloid-like tau induced by 0.1%, 0.05% or 0.01% formaldehyde each produced a significant decrease on the proportion of apoptotic cells (23%, 22% and 15%, Fig.  9D-F, respectively).

Discussion
Methanol is an ocular toxicant that causes visual dysfunction often leading to blindness after acute exposure. The physiological and biochemical changes responsible for this toxicity are poorly understood [44]. According to a recent report, humans are uniquely sensitive to the toxicity of methanol, as they have limited capacity to oxidize and detoxify formic acid. Thus, the toxicity of methanol in humans is characterized by formic acidaemia, metabolic Tau deposits were imaged by transmission electron micros-copy and atomic force microscopy  A acidosis, blindness or serious visual impairment and mild central nervous system depression or even death [1,44]. This view is based on the following two observations: (1) methanol is metabolized to formaldehyde in liver cells and also in neurons [24][25][26][27]45], although it is very rapidly converted to formate; (2) SSAO-mediated generation of formaldehyde can induce protein (i.e. β-amyloid) crosslinkage, deposition and subsequently plaque formation in Alzheimer's disease [29][30][31].
In recent years, however, formaldehyde has been found to be a neurotoxic molecule and to damage the prefrontal cortex of rats including the hippocampus [46,47]. These results demonstrate the formaldehyde-induced neurotoxicity to neurons. Our studies show that formaldehyde induces neuronal tau to aggregate. Here, we show that amyloid-like tau induces apoptosis of SH-SY5Y and hippocampal cells. In fact, chemically, formaldehyde reacts with thiol (our unpublished data) and amino groups instantly, while misfolding of neuronal tau is a subsequent event. This suggests that amyloid-like tau may be involved in methanol toxicosis, particularly the chronic damage to neurons.
The microtubule associated protein tau plays an important role in maintenance of the cytoskeleton. It promotes and maintains assembly of microtubules, which are required for axonal morphogenesis and transport [48]. In recent years, PHFs, formed by misfolding of tau, were found to be the main component of neurofibrillary tangles involved in neurodegeneration, such as in Alzheimer's disease. PHF-tau is not only commonly found in Alzheimer's brain, but is also induced by simple incubation of native tau with some glycosaminoglycans, for instance heparin, in vitro [5,32]. Here, we found that formaldehyde-treated tau forms amyloid-like aggregates, although not necessarily PHFs. Certainly, under the conditions used, self-aggregated tau showed certain differ-ences in structure compared with the aggregates induced by exposure to formaldehyde. (1) Congo Red assays showed that the dye absorbance increased by 16% after incubation with formaldehyde-treated tau, and the absorbance increase was accompanied by a red-shift to 510 nm. Similarly, when formaldehyde-treated tau was added to ThT, a 4-fold increase in the emission intensity and the emission maximum shifted to 482 nm were observed. In contrast, self-aggregated tau induced little change in the spectra of this amyloid-specific dye. (2) Electron microscopy showed that formaldehyde-treated tau had the appearance of granular amyloid-like aggregates, with the diameters in the range of 20-100 nm, unlike fibrillary structures in PHF-tau. (3) The results observed by AFM further confirmed the presence of globular aggregates under the same conditions. The results suggest that formaldehyde promotes the formation of amyloid-like aggregates, which may represent a variant of tau amyloid-like structure.
Recently, Kuret and colleagues described a tau assembly pathway in which anionic inducers, for instance arachidonic acid (AA), favor a shift in the equilibrium between unfolded and filamentous tau species. The microtubule binding function of tau is lost and tau protein accumulates in a partially folded, ThT-positive intermediate which then self-aggregates into a hydrophobic nucleus (as detected by fluorescent ANS), before the filament nucleus elongates to form full fibrils [37,39]. In contrast, formaldehyde-tau was not observed to elongate into filaments on the experimental timescale used in this paper. However, as formaldehyde is not an anionic inducer, it is not surprising that different characteristics are observed between formaldehyde-and AA-induced tau aggregates.
As shown above, formaldehyde reacted with the amino groups of tau, as demonstrated by the OPT test. Reaction with formaldehyde is known to eliminate positive (NH 2 ) Contrast microscope image of cells treated with amyloid-like tau A B C D groups and to increase the net negativity of a protein [49], which may lead to conformational changes in protein tau. Unlike tau, however, formaldehyde at the low concentrations used here did not induce any detectable degree of aggregation or conformation change in BSA. According to Schweers et al. (1994) [48], the conformation of native tau features a "worm-like" or a "denatured-like" structure, leaving ε-amino groups of Lys exposed to the exterior of the tau molecule, which would allow formaldehyde to interact with the amino groups of tau. Furthermore, it has been reported that neuronal tau is prone to aggregation when incubated at 37°C or room temperature for over 10 h [5,32]. On the other hand, in BSA, a globular protein, not all of the ε-amino groups are accessible for reaction with formaldehyde. As a crosslinking agent for globular proteins, formaldehyde is not so particularly efficient. Glutaraldehyde is commonly used because the linker region is long enough to bridge two protein molecules. The fact that neuronal tau is prone to aggregate when exposed to low concentrations of formaldehyde, probably reflects the unfolded nature of its native conformation.
Khlistunova and colleagues found that the repeat domains of intracelluar tau could aggregate and were toxic to neuronal cells. The degree of tau aggregation and toxicity depends on the propensity to form β-structure [15,38,50,51]. In the present study, we found that extracelluar tau aggregates can induce neuronal cell apoptosis, similar to the results obtained with extracelluar amyloid or α-synuclein [7,8,43,52,53]. This suggests that structures enriched in β-sheet are important for amyloid-like protein aggregation and neurotoxicity. Hence intracelluar amyloid-like proteins can form neurotoxic aggregates in vitro.
In our experiments, a low concentration of formaldehyde induced recombinant tau to aggregate into cytotoxic amyloid-like granular aggregates, providing a new potential mechanism for tauopathies. However, our work provides an effect on protein tau aggregation in vitro of low concentrations of formaldehyde. For an in vivo environment where many other biochemical and biophysical factors exist and interact with each other, further investigation needs to becarried out.

Conclusion
Here we investigate the effect of low concentrations of formaldehyde on protein misfolding and aggregation. We found that unlike the typical globular protein BSA, the natively-unfolded structure of human neuronal tau was induced to misfold and aggregate in the presence of 0.01% formaldehyde, leading to formation of amyloidlike deposits that appeared as densely staining granules by electron and atomic force microscopy, and bound the amyloid-specific dyes thioflavin T and Congo Red. After removal of the formaldehyde, the amyloid-like aggregates of tau were found to induce apoptosis in the neurotypic SH-SY5Y cells and in rat hippocampal cells, as observed by Hoechst 33258 staining, assay of caspase-3 activity, and flow cytometry using Annexin V and Propidium Iodide staining. Control cells incubated with formaldehyde alone, or with tau aggregates formed in the presence of acetaldehyde or in the absence of additives (and which did not show appreciable binding of thioflavin T or Congo Red), did not show signs of apoptosis. Further experiments showed that Congo Red specifically attenuated the caspase-3 activity induced by amyloid-like deposits of tau. The results suggest that low concentrations of

Expression and purification of recombinant htau-40 and microtubule binding assay
Neuronal tau, overexpressed in E. coli, was purified as described previously [54][55][56]. Briefly, the bacterial cells were homogenized with a sonicator, boiled at 100°C and the protein was purified using Sepharose-Q, Sepharose-SP and Sephdex-50 chromatography columns. The concentration of recombinant tau was determined spectrophotometrically by measuring the absorbance at 280 nm [57] and the protein appeared as a single band in SDS-PAGE after purification. Assay of tau was performed with tubulin according to Alonso et al. [3]. Tau (4 μM) was mixed at 4°C with purified porcine brain tubulin (1 mg/ml) and 1 mM GTP, all in polymerization buffer (100 mM MES, pH 6.7, 1 mM EGTA, and 1 mM MgCl 2 ) in a final volume of 500 μl. Once mixed, the samples were pipetted into quartz microcuvettes (1 ml), equilibrated at 37°C, and the absorbance at 350 nm was measured on a Hitachi U-2010 spectrophotometer. The specific activity of tau was the same as that described by Goedert et al. (1990) [58].

Light scattering and electrophoresis
For protein aggregation experiments, tau (1.2 μM for light scattering or 20 μM for electrophoresis) was incubated at 37°C with different concentrations of formaldehyde in 100 mM phosphate buffer (pH 7.2) for 24 h to allow the reaction to reach completion. Acetaldehyde and BSA were used as controls. For time course analysis, aliquots were taken at different time intervals during incubation of tau with or without formaldehyde. Light scattering was measured in a Hitachi F-4500 fluorescence spectrophotometer (slits: Em = 5.0 nm and Ex = 5.0 nm) with excitation at 480 nm. Kinetic data were analyzed according to Tsou (1965) [59]. Electrophoresis equipment was from Bio-Rad.

Fluorescence and CD measurements
To test conformational changes, ANS was employed to detect whether the aggregates of formaldehyde-treated tau had exposed hydrophobic surface area. The fluorescence at 480 nm of ANS after excitation at 350 nm for different concentrations of tau was determined at room temperature using a Hitachi F-4500 spectrofluorometer. CD spectra were recorded using a Jasco J-720 CD spectrometer. The spectra were measured in 1-mm pathlength quartz cuvettes, and data were collected from 195 nm to 250 nm at 0.5-nm intervals. The samples were all in 100 mM sodium phosphate buffer (pH 7.2). The bandwidth was set at 1.5 nm (37°C). The spectrum baselines were corrected using the spectrum for the buffer measured under identical conditions.

OPT modification
Neuronal tau (final concentration 0.1 μM) was resuspended in 50 mM phosphate buffer containing OPT (molar ratio: reagent/protein = 20/1) in the presence of formaldehyde at different concentrations at 37°C for 120 min. The fluorescence (Ex340 nm/Em455 nm) was then measured (slits: Em = 5.0 nm and Ex = 5.0 nm). Under the same conditions, tau was resuspended in 0.005% formaldehyde and 2 μM OPT; and aliquots were taken to measure the fluorescence at different time intervals. The data were analyzed according to Tsou (1965) [59].

Congo red and thioflavin T binding assays
The assay [36] was performed by adding a freshly prepared stock solution of Congo red in 100 mM potassium phosphate (pH 7.2) to tau samples (2 μM) to give a final Congo Red concentration of 5 μM. The ThT assay was performed by adding a freshly prepared solution of ThT in the phosphate to tau samples to give a final ThT concentration of 10 μM [35]. The protein absorbance contribution was subtracted from the spectra.

Electron microscopy
For electron microscopy, tau (40 μM) was incubated with 0.1% formaldehyde at 37°C for 24 h. The incubated samples were loaded on a carbon-coated grid for 2 min, stained with 2% (w/v) uranyl acetate for 1 min, and then dehydrated through a graded water-ethanol series. Samples were visualized under a JEOL JEM-100CX electron microscope. Tau alone (40 μM) was incubated under the same conditions in the presence of heparin (1 mg/ml) as a control.

Atomic force microscopy
Neuronal tau (final concentration 10 μM) was incubated at 37°C with formaldehyde (0.05%) in 25 mM phosphate buffer (pH 7.2) over night. Then protein solution was diluted using phosphate buffer and 3 μl of the sample (10 ng tau protein) was dropped onto the mica surface and left for 5 min at room temperature before drying with nitrogen gas. The mica diaphragm was rinsed with ultrapurified water 20 times and dried down with nitrogen gas before observation under the atomic force microscope (Mutiplemode-I, Digital Instruments). The horizontal diameter at half height of a particle (globular protein) was measured and the data were analyzed using the Nanoscope 6.11r1 software.