Journal of Alzheimer’s Disease 35 (2013) 319–334 319

DOI 10.3233/JAD-121765
IOS Press

Molecular Mechanism of Tau Aggregation

Induced by Anionic and Cationic Dyes
Karla I. Lira-De Leóna , Ponciano Garcı́a-Gutiérrezb , Iris N. Serratosc , Marianela Palomera-Cárdenasd ,
Marı́a del P. Figueroa-Coronaa , Victoria Campos-Peñae and Marco A. Meraz-Rı́osa,∗
a Departamento de Biomedicina Molecular, Centro de Investigación y de Estudios Avanzados del Instituto Politécnico

Nacional CINVESTAV-IPN, México, D.F.

b Área de Biofisicoquı́mica, Departamento de Quı́mica, Universidad Autónoma Metropolitana-Iztapalapa,

México, D.F.
c Área de Fisicoquı́mica de Superficies, Departamento de Quı́mica, Universidad Autónoma Metropolitana-

Iztapalapa, México, D.F.

d Instituto Tecnológico de Sonora, Sonora, México
e Laboratorio Experimental de Enfermedades Neurodegenerativas, Instituto Nacional de Neurologı́a y Neurocirugı́a

Manuel Velasco Suárez, México, D.F.

Handling Associate Editor: Emilio Di Maria

Accepted 23 January 2013

Abstract. Abnormal tau filaments are a hallmark of Alzheimer’s disease. Anionic dyes such as Congo Red, Thiazine Red, and
Thioflavin S are able to induce tau fibrillization in vitro. SH-SY5Y cells were incubated with each dye for seven days leading
to intracellular aggregates of tau protein, with different morphological characteristics. Interestingly, these tau aggregates were
not observed when the Methylene Blue dye was added to the cell culture. In order to investigate the molecular mechanisms
underlying this phenomenon, we developed a computational model for the interaction of the tau paired helical filament (PHF)
core with every dye by docking analysis. The polar/electrostatic and nonpolar contribution to the free binding energy in the tau
PHF core-anionic dye interaction was determined. We found that the tau PHF core can generate a positive net charge within the
binding site localized at residuesLys311 and Lys340 (numbering according to the longest isoform hTau40). These residues are
important for the binding affinity of the negative charges present in the anionic dyes causing an electrostatic environment that
stabilizes the complex. Tau PHF core protofibril-Congo Red interaction has a stronger binding affinity compared to Thiazine
Red or Thioflavin S. By contrast, the cationic dye Methylene Blue does not bind to nor stabilize the tau PHF core protofibrils.
These results characterize the driving forces responsible for the binding of tau to anionic dyes leading to their self-aggregation
and suggest that Methylene Blue may act as a destabilizing agent of tau aggregates.

Keywords: Alzheimer’s disease, Congo Red, docking, Methylene Blue, SH-SY5Y cells, tau aggregation, tauopathies, Thiazine
Red, Thiofavin S

INTRODUCTION lesions. Because Alzheimer’s disease (AD) is the

Tauopathies are a group of neurodegenerative
diseases characterized by the aggregation of the
microtubule-associated tau protein into filamentous
∗ Correspondence to: Dr. Marco A. Meraz-Rı́os, CINVESTAV-
IPN, Av. I.P.N. 2508 C.P. 07360, México, D.F. E-mail: mmeraz@
cinvestav.mx.
most common of these pathologies, studies of the
pathological process of tau aggregation are of great
interest [1, 2]. Tau protein is the main component in
neurofibrillary tangles (NFTs) and the amount of these
aggregates correlates with cognitive impairment [3].
This association is based on tau filament formation
resulting in cytoskeletal disorganization, one of the

ISSN 1387-2877/13/$27.50 © 2013 – IOS Press and the authors. All rights reserved
320 K.I. Lira-De León et al. / Tau Aggregation Induced by Dyes
characteristic features of degenerating neurons [4].
However, under physiological conditions, tau is a
highly soluble protein with a poor secondary structure
that polymerizes weakly. Studies on NFTs have con-
firmed that tau aggregates display ␤-sheet structures.
This conformational change is possible because tau
protein contains two motifs which have a regional
trend to form ␤-sheet structures, flanked by the second
(275 VQIINK280 ) and the third (306 VQIVYK311 )
repeats of the microtubule-binding domain (MTBD)
[5–7]. This process can be triggered by exogenous
fibrillization promoters such as polyanions, anionic
micelles, and anionic dyes [5]. The mechanism
proposed for tau conversion from a natively unfolded
protein into a filament form involves an allosteric
transition intermediate in which anionic dyes such as
Congo Red (CR), Thiazine Red (TR), and Thioflavin S
(TS) are capable of inducing tau fibrillization in vitro.
The most potent and effective inducer is CR since
treatment with a low concentration (10 ␮M) already
induces a large numbers of filaments. TS is able to
produce a large number of small filaments but requires
a 10-fold higher concentration. By contrast, treatment
with high concentrations of TR only yields few fila-
ments of relatively long length [8]. However, only CR
at concentrations of 10 ␮M has been reported to induce
tau aggregation in the human Embryonic Kidney-293T
cell line (HEK-293T) which is accompanied by a
significant loss of cell viability. However, the cell
type used in this study was not of neuronal origin and
showed resistance to the uptake of both TS and TR [9].
In order to prevent tau aggregation, several groups
have been interested in the use of certain drugs. For
example, Methylene Blue (MB) is a member of the
phenothiazine family that crosses the blood-brain bar-
rier. That has been shown in vivo to enhance the
functionality of the proteasome and to decrease AD
pathology. It has also been observed in an animal model
of AD that MB has the ability to rescue impaired
learning and memory [10]. Phase II clinical trials in
AD already demonstrated a significant improvement
of cognitive functions in patients treated with MB [11].
In this report, we used the human neuronal cell
line SH-SY5Y to test the capacity of the anionic
dyes CR, TR, and TS to promote tau aggregation.
Additionally, we analyzed the inhibitory effects of
MB on tau aggregation. To characterize the molecular
interaction between tau and anionic dyes, we computa-
tionally modeled tau paired helical filament (PHF) core
protofibrils and carried out extensive docking simula-
tions with CR, TR, TS, and MB. This allowed us to
determine the polar/electrostatic (
Gp ) and nonpolar
(
Gnp ) contributions to the free binding energy in the
tau PHF core-anionic dye interaction.
We found that the tau PHF core protofibrils contain
positive charges at lysine residues that interact with
the negatively charged anionic dyes to provide elec-
trostatic stability of these structures, which does not
occur with MB.
MATERIALS AND METHODS
Cell culture
SH-SY5Y cells were grown in 100-mm diameter
dishes (Costar, Corning Incorporated, NY, USA) at
80% confluence in Advanced Dulbecco’s modified
Eagle’s medium (A-DMEM) (Invitrogen/Gibco,
Carlsbad, CA, USA) supplemented with 10% heat-
inactivated fetal bovine serum (Invitrogen/Gibco),
100 U/mL penicillin, and 100 ␮g/mL streptomycin
(BIOSOURCE Inc., Rockville, MD, USA) and
maintained at 37◦ C in 95% air and 5% CO2 in a
humidified chamber.
Dye solutions
CR (GOLDEN BELL reactivos I.C., Jalisco,
México) and TR (Sigma, St. Louis, MO, USA) were
stored as a 10 mM stock solution in distilled water. TS
(Sigma) was stored as a 10 mM stock solution in a 1:1
mixture of ethanol and distilled water. MB (Sigma) was
stored as a 10 mM stock solution in distilled water [12].
All solutions were sterile-filtered using a 0.22 ␮m filter
[9]. Anionic dyes were stored at 4◦ C and MB solution
at room temperature (RT) for one month.
Anionic dye treatment
SH-SY5Y cells were grown in monolayers on
twelve-well plastic plates (Costar, Corning Incor-
porated) in a density of 300,000 cells per dish
and maintained with the respective dye at different
concentrations (0, 5, 10, 15, 30, 60, 90, and 100 ␮M)
for seven days. Every third day half of the total
medium volume was replaced by fresh culture
medium with the respective concentration of dye. For
immunocytochemistry analysis, cells were cultured on
poly-D-lysine-coated (10 ␮g/mL) (ICN, Cleveland,
Ohio, USA) glass cover slips.
Methylene Blue treatment
SH-SY5Y cells were grown as described above
except for the last day of incubation the cells were
 K.I. Lira-De León et al. / Tau Aggregation Induced by Dyes
treated with MB at different concentrations (0, 1, 2,
2.5, 5, 10, 20, 30, 40, 50, 80, and 100 ␮M) for 24 h.
Immunocytochemistry
Cells were fixed with 4% paraformaldehyde (EMS,
Morris, Washington, USA) for 30 min and permeabi-
lized with 0.2% Triton X-100 (Sigma) by 3 washes
of 10 min each time. Following blocking with 1%
bovine serum albumin fraction V (BSA) (Sigma) for
30 min, a polyclonal rabbit antibody against tau Ab3
(Neo Markers/Lab Vision, Fremont, CA, USA) was
applied overnight at 4◦ C. Cells were then incubated
with goat anti-rabbit secondary antibody conjugated
to Alexa 488 (1:1000) or Alexa 594 (1:1000) (Invitro-
gen/Molecular probes) for 1 h. Secondary antibodies
only served as negative controls. Slides were mounted
using Vectashield (Vector Laboratories) and visualized
using a confocal laser microscope (Olympus IX71)
and FluoView FB300 3.2 browser imaging software
(Olympus America Inc., Center Valley, PA, USA).
Flow cytometry
Cell cultures were detached with 0.5% trypsin-
EDTA (Invitrogen/Gibco, Grand Island, NY, USA),
washed, resuspended in PBA (PBS 1X, 0.1% BSA,
0.01% sodium azide) and treated with LIVE/DEAD
Viability/Cytotoxicity kit (Molecular Probes, Eugene,
OR USA). Cells were distributed into 96-well
U-bottomed microtiter plates (Costar, Corning Incor-
porated) and resuspended in FPA (Formaldehyde 1%,
PBS 1X, 0.01% sodium azide) prior to flow cytometry
using a FACS caliber flow cytometer (Becton Dickin-
son, Mountain View, CA, USA). Results were analyzed
using Cell Quest software (Dickinson). Untreated cells
served as controls.
Sarcosyl insoluble extraction
Cell cultures were detached with 0.5% trypsin-
EDTA (Invitrogen/Gibco), washed, homogenized in 10
volumes of buffer (10 mM Tris, 0.8 M NaCl, 1 mM
EGTA and 10% sucrose, pH 7.4), and centrifuged
at 27200× g for 30 min [13]. The supernatant was
adjusted to 1% (w/v) N-lauroylsarcosine and incubated
1 h at RT. After incubation, the supernatant was spun at
123000× g for 1 h at RT. The pellet was resuspended
in a small volume (20 ␮l) of phosphate buffered saline
(PBS). Additionally, an aliquot of the supernatant of
equal volume as the pellet was taken and finally the
samples were subjected to western blot.
321
Tau aggregation membrane filter assay
Cell cultures were detached with 0.5% trypsin-
EDTA (Invitrogen/Gibco), washed, and homogenized
in 50 ␮l of buffer (20 mM Tris, 0.15 M NaCl, 1 mM
EDTA, and 1% Triton X-100, pH 7.5) with protease
and phosphatase inhibitors and centrifuged at 27200×
g for 30 min. The samples were filtered through
0.22 ␮m PVDF membranes [14]. The resultant mem-
branes were blocked in 5% nonfat dry milk dissolved
in blocking buffer (10 mM Tris-HCl, pH 7.4, 150 mM
NaCl) for 1 h, and then incubated with tau primary anti-
body [mouse monoclonal DC25 to MTBD-minimal
epitope: 347–353 aa (rPeptide, LLC), mouse mono-
clonal T46 to C-terminal- minimal epitope: 404–441
aa (Santa Cruz Biotech), rabbit polyclonal H150 to
N-terminal- minimal epitope: 1–150 aa (Santa Cruz
Biotech)] or ␣-actin or ␣-tubulin primary antibody
overnight. Membranes were washed three times in
Tween buffer (10 mM Tris-HCl, pH 7.4, 150 mM
NaCl, 1% Tween 20), then incubated with HRP-linked
secondary antibody for 1 h. The membranes were
washed three times in Tween buffer. Membranes were
developed using the ECL plus Western Blotting Analy-
sis System (Amersham Bioscience, Buckinghamshire,
UK).
Electron microscopy
Pellets obtained from Sarcosyl extraction were
analyzed by electron microscopy. Aliquots were coun-
terstained with uranyl acetate and lead citrate. Images
were viewed in transmission electron microscope oper-
ated at 60-40 kV [6, 15].
Computational docking between tau PHF core
protofibrils and dyes
Molecular modeling, electric partial-charge assigna-
tion, ligand complex, energy minimizations, docking,
and visualization were performed with the Molecu-
lar Operating Environment (MOE) (version 2007.9)
[16] package unless otherwise stated. Energy mini-
mizations were carried out until an RMSD (root mean
square deviation) lower than 0.05 was obtained using
CHARMM27, an all-atom force field parameterized
for proteins, or MMFF94x, parameterized for small
organic molecules in medicinal chemistry.
Modeling of tau PHF core
Five 3D models of tau PHF core sequence were built
based on multiple-threading alignments by LOMETS
322 K.I. Lira-De León et al. / Tau Aggregation Induced by Dyes
(http://zhanglab.ccmb.med.umich.edu/LOMETS/) and
iterative TASSER assembly simulations using the
I-TASSER server (http://zhanglab.ccmb.med.umich.
edu/I-TASSER/) and refined by energy minimization
(MOE EnergyMinimize) up to RMSD of 0.1 using the
CHARMM27 force field. The best model (Model 1)
was selected on basis of their low potential energy and
Ramachandran plot and was used to construct tau-tau
PHF core protofibril.
Modeling of tau PHF core protofibril
The assembly system consisted of a 16-peptide
protofibrillar oligomer. This 16-peptide protofibril-
lar oligomer was constructed from two anti-parallel
␤-strands present in tau PHF core Model 1 (i.e.,
306 VQIVYK311 and 335 GQVEVKS341 third and fourth
repeats, respectively). According to Mukrasch, it has
been reported that these repeats are known to be
involved in the abnormal aggregation of tau [17, 18].
The 16 peptides were arranged in a double-layered ␤-
sheet structure with each sheet formed from parallel
segments (one segment consist of anti-parallel beta-
strands shown in Model 1) stacked in register. The two
␤-sheets were related to one another by an S2 sym-
metry axis (i.e., a 180◦ rotation followed by reflection
through the plane perpendicular to the rotation axis).
Side chains protruding from the two sheets form a
zipper bonding the sheets. Within each sheet, every
segment is bound to its two neighboring segments
through stacks of both backbone and side chain hydro-
gen bonds. Our constructed protofibrillar oligomer is
based on the NMR model of Luührs et al. [19] for the
A␤17–42 structure on recent experimental data [17, 18]
that provides some details of the secondary structure of
the tau protein and its X-ray interaction with orange-G
[20]. A side chain of residues in the protofibril was sub-
mitted to energy minimization using the CHARMM27
force field, while the rest of the structure was fixed.
Computational docking analysis
Three-dimensional structures of the dye molecules
were generated using MOE Buid and submitted to
energy minimization using the MMFF94x force field.
Additional 3-D conformations of each dye were gen-
erated through MOE ConfomerSearch using a cut-off
conformational energy of 7 kcal/mol from the mini-
mum energy structure of each compound. Conformers
of dye molecules were docked to the rigid tau PHF core
protofibril structure using MOE Dock and default con-
figurations. The binding simulation between each dye
conformer and tau PHF core protofibril was performed
including the complete surface of the protofibril as well
as the interface region. The top 50 docking scores of
complexes consisting of tau PHF core protofibril and
dye conformer (15,000 different binding orientations
on potential binding sites were proved and evaluated
for each conformer from CR, TR, TS, and MB) were
further refined and optimized using energy minimiza-
tion. The best tau PHF core protofibril-dye molecule
complex based on lowest ligand-protein energy inter-
actions and intermolecular interactions was chosen as
the final structure for each dye molecule.
Theoretical determinations: Polar and nonpolar
contributions to the free binding energy
The solvation energies (
Gsolv ) of tau PHF
protofibril-dye molecule complexes were calcu-
lated using the numerical solution to the nonlinear
Poisson–Boltzmann equation implemented in the pro-
gram APBS (Adaptive Poisson–Boltzmann Solver)
[21]. Dielectric constants of 78 and 4 were used for
water and protein, respectively. These energies were
calculated at 298.15 K and at 150 mM NaCl. Parame-
ters were taken from the PDB2PQR server [22], which
converts PDB files into PQR files: ionic radii and
atomic charges were assigned from the CHARMM
forcefield [23] and terminal amino and carboxyl groups
were taken as neutral. PROPKA [24] was used to
assign the protonation state of ionizable residues at pH
7.4. For anionic dyes, atomic charges were assigned
from the forcefield MMFF94x implemented in pro-
gram MOE.
Using a thermodynamics cycle for the calculation
of the electrostatic free energy of solvation contribu-
tion to the binding free energy as described by Baker
[21]:


Gsolv =
Gsystem −
Greference (1)
=
GTau PHF core protofibril-dye complex
(2)
−
GTau PHF core protofibril −
Gdye
The Coulombic energies (
Gcoul ) of tau protofibril-
anionic dye complexes were calculated with a program
Coulomb within APBS accessory [21, 25] according
to:

Gcoul = GTau PHF core protofibril-dye complex
(3)
− GTau PHF core protofibril − Gdye
 K.I. Lira-De León et al. / Tau Aggregation Induced by Dyes
The polar energy was calculated using the following
expression:

Gp =
Gsolv +
Gcoul (4)
The nonpolar contribution to the binding energy was
estimated by multiplying the change in solvent accessi-
ble surface (
ASA) upon binding by 5 cal mol−1 Å−2
[26] leading to an additional term (
Gnp ). The calcu-
lations of ASA were done using the program Visual
Molecular Dynamics (VMD) implying a probe radius
of 1.4 Å and the atomic coordinates of PDB files tau
PHF core protofibril-dye molecule:

Gnp =␥ (ASATau PHF core protofibril-dye complex
(5)
− ASATau PHF core protofibril − ASAdye )
where coefficient ␥ is an effective microscopic inter-
facial tension (5 cal mol−1 Å−2 )
In summary, the calculated energy (
Gcalc ) is deter-
mined by:

Gcalc =
Gsolv +
Gcoul +
Gnp (6)
RESULTS
Effect of anionic dyes on SH-SY5Y cells
Treatment of SH-SY5Y cells with the dyes CR, TS
and TR promotes tau aggregation. Thus, SH-SY5Y
cells were treated with the respective dyes at different
concentrations (0, 5, 10, 15, 30, 60, 90, and 100 ␮M)
for seven days. Internalization of dye was measured by
flow cytometry as shown in Fig. 1A. The percentage
of positive cells was exact value 55%. CR showed a
particularly strong response, since at the lowest con-
centration there was already maximal uptake. On the
other hand, both TR and TS showed a dose-dependent
response with a similar percentage of positive cells at
60 ␮M for TS and 100 ␮M for TR (Fig. 1A). SH-SY5Y
cells treated with dyes showed severe damage in their
integrity and cell viability compared with untreated
cells (Fig. 1B). TR was the most aggressive dye with
only 20–30% of viable cells. TS treatment showed
average toxicity (40–55%) whereas CR-treatment did
not show significant cytotoxic effects (92–99%) simi-
lar to untreated cells.
Tau can interact with anionic dyes in SH-SY5Y
cells
Cells treated with anionic dyes at a concentration of
15 ␮M still showed a normal morphology comparable
323
with untreated control cells. Using an anti-tau antibody,
we showed by immunocytochemistry co-localization
between the respective dye and tau in the cytoplasm
of SH-SY5Y cells (Fig. 2). In treated cells, tau pro-
tein showed a similar distribution as in cells from AD
patients resembling a deposit which implies the appear-
ance of protein aggregates.
Tau aggregates induced by anionic dyes in
SH-SY5Y cells
Formation of tau aggregates was investigated in
a membrane filtration assay. Aggregates unable to
pass through the pore membrane were detected by
immunostaining of the membrane with anti-tau anti-
bodies (DC25 and T46). Cells treated with 100 ␮M CR,
100 ␮M TR, and 90 ␮M TS for seven days were col-
lected, resuspended in lysis buffer, and cell lysates were
analyzed in the membrane filtration assay (Fig. 3A).
All membranes showed positive staining with both
antibodies for all treated cells indicating the presence
of tau aggregates composed, at least, of the MTBD
and the C-terminal region of tau, epitopes that are
recognized by the antibodies (Fig. 3B, C). By con-
trast, no positive signal was detected in untreated
SH-SY5Y cells. A positive signal with the H150
antibody was exclusive for TR treatment (Fig. 3D)
suggesting that in CR and TS treatments, tau aggre-
gates are composed of N-terminally truncated tau.
Importantly, no cytoskeletal proteins such as actin
or tubulin could be detected in these aggregates,
(Fig. 3E, F).
Tau aggregates morphology induced by anionic
dyes
In order to evaluate the morphology of these aggre-
gates, we prepared pellets by sarcosyl extractions for
analysis by electron microscopy. The morphology of
the obtained filaments was different and specific for
each dye (Fig. 4). For instance, the filaments formed
with CR are robust and amorphous (Fig. 4B) while TR
filaments are short and straight (Fig. 4C). Interestingly
TS filaments showed a like-paired helical morphology
and long length (Fig. 4D).
Effect of MB treatment in SH-SY5Y cells
Tau aggregation induced by planar dyes in SH-
SY5Y cells could be used as a model of tau
polymerization. Moreover, it is possible to test drugs
324 K.I. Lira-De León et al. / Tau Aggregation Induced by Dyes

Fig. 1. Anionic dye uptake by and viability of SH-SY5Y cells. A) Percentage of cells that had internalized the dye and (B) percentage of
viable cells as determined by the LIVE/DEAD Viability/Cytotoxicity kit as described in Materials and Methods. The percent of positive cells
corresponds to the dye uptake in the first case and staining with LIVE/DEAD Viability/Cytotoxicity kit in the second case. Data are mean ± SEM
of three independent experiments *p < 0.05 by Student’s t test compared to untreated cells.

that inhibit or destabilize the abnormal aggregation of MB destabilizes tau aggregates

tau. In this respect, we tested the destabilizing capac-
ity of MB on tau aggregates. To this end, SH-SY5Y
cells were treated with different MB concentrations
(0, 1, 2, 2.5, 5, 10, 20, 30, 40, 50, 80, and 100 ␮M)
for 24 h. We found that 40 ␮M was the concentration
with the highest percentages of positive (blue) cells
(Fig. 5A). Analyzing the morphology of cells treated
with 40 ␮M of MB, we found a healthy morphology
similar to untreated cells (Fig. 5B). However, treatment
with 80 ␮M MB caused severe damage to the cells as
shown in Fig. 5C.
The MB destabilization assay for tau aggregation
was conducted as follows: SH-SY5Y cells were treated
with anionic dyes as describe above and for the last
24 h MB was added to the culture medium (Fig. 5D).
Strikingly, we found a concentration-dependent reduc-
tion in the number of cells containing tau aggregates.
The average destabilizing MB concentration (50%
of cells) was different for each anionic dye: 10 ␮M
for CR, 20 ␮M for TR, and 2.5 ␮M for TS. These
results suggest that MB can more efficiently inhibit
 K.I. Lira-De León et al. / Tau Aggregation Induced by Dyes 325

Fig. 2. Tau colocalizes with anionic dyes in SH-SY5Y cells as determined byconfocal laser microscopy. A, D) SH-SY5Y control cells.
B, C, E) SH-SY5Y cells treated with 15 ␮M of CR, TR, and TS respectively. Arrows indicate regions of colocalization. Scale bar = 20 ␮m.

tau aggregation induced by CR and TS compared to TR
treatment. To observe tau aggregation by immunoflu-
orescence, we also used the filter assay in cells treated
with anionic dyes (Fig. 6A) in combination with MB
(Fig. 6C). As expected, tau immunostaining was nega-
tive in all co-treatments (Fig. 6D) compared with cells
without MB treatment (Fig. 6B). Thus, our results indi-
cate that the stability of tau aggregates is reduced by
MB treatment.

Ultrastructure analysis of tau aggregates induced

by anionic dyes on SH-SY5Y cells

Electron microscopy analysis of pellets obtained by

Fig. 3. Tau aggregates are induced by anionic dyes in SH-SY5Y
cells. A) Membrane filtration assay of SH-SY5Y cells treated
with 100 ␮M CR, 100 ␮M TR, and 90 ␮M TS. B–D) Tau immunos-
taining with specific antibodies: MTBD, C-terminal, and N-terminal
respectively. E, F) Immunostaining of actin and tubulin proteins.
sarcosyl extractions showed the absence of filaments in
all cotreatments with MB (Fig. 7B–E) compared with
TS treated cells alone (Fig. 7A), which showed the
characteristic filament structures for tau aggregation
reported previously.
Modeling tau PHF core and protofibril
In order to understand the interaction between tau
PHF core-anionic dyes and MB, we performed docking
326 K.I. Lira-De León et al. / Tau Aggregation Induced by Dyes
Fig. 4. Morphology of tau aggregates induced by anionic dyes in SH-SY5Y cells as analyzed by electron microscopy. A) Control. B) Treatment
with 100 ␮M CR. C) Treatment with 100 ␮M TR. D) Treatment with 90 ␮M TS. Scale bar = 100 nm.
simulations. Because the domain of tau that has a ten-
dency to form ␤-pleated structures sheets is the PHF
core, we used this fragment in order to model the tau
PHF core and the protofibril.
Five models of tau PHF core protein have been
constructed (see Material and Methods for more
details). All tau PHF core models have very little
secondary structures. Two ␤-strands forming an
anti-parallel ␤-sheet from repeats 306 VQIVYK311 and
335 GQVEVKS341 are present in all models which are
stabilized by hydrogen bonds. The rest of the amino
acid sequence is unstructured. Structure superposition
for all models shows RMSD values between 2.5 and
5.3 Å considering all atoms, but the value drops to
0.7–1.2 Å if considering only the C␣ atoms. The
Ramachandran plot for Model 1 (Fig. 8) shows the
residues at the ␤-sheet region in allowed regions and
this model has the lowest potential energy compared
to the other models making Model 1 the best to be
used for modeling the tau PHF core protofibril system
using docking studies. It has been suggested that
the motifs 275 VQIINK280 , in the second repeat, and
306 VQIVYK311 , in the third repeat of tau, play a key
role in the formation of PHF [27]. An X-ray study of
306 VQIVYK31 microcrystals revealed a cross-␤ spine
structure consisting of ␤-sheet pairs with a “dry steric
zipper” organization at the sheet-sheet interface [28].
In our model, we used a fragment constituted by the
␤-sheet observed in Model 1 to build the tau PHF
core protofibril and two ␤-sheet form the steric zipper
(Fig. 9). The dimensions of the tau PHF core protofib-
rillar model are 36 × 22 × 25 Å along the ␤-sheet
extension, ␤-strand, and ␤-sheet stacking (perpendi-
cular to the ␤-sheet surface) directions, respectively.
Computational docking
To simulate the ligand flexibility in our docking
studies, we generated a set of low-energy conformers
for each dye molecule. About two hundred of con-
formers with energies ≤7 kcal/mol with respect to the
lowest minimum conformation for each dye molecule
were built (209, 195, 182, and 217 conformers for CR,
TR, TS, and MB, respectively). The previously gener-
ated conformers for each dye molecule were scanned
against all atoms in the tau PHF core protofibril model
(overall surface and interface region) using 15,000 test
poses for each conformer in order to find the best bind-
ing pose. Then, energy minimization of complexes
with the best docking scores for each dye was carried
out to define the dye orientation and to identify the
interacting amino acid residue within the tau PHF core
protofibril.
 K.I. Lira-De León et al. / Tau Aggregation Induced by Dyes 327

Fig. 5. MB treatment inhibits tau aggregation. A) SH-SY5Y cells treated with different concentrations of MB for 24 h. B) Morphology of cells
treated with MB. C) Analysis of cell morphology after MB treatment. D) SH-SY5Y cells treated with 100 ␮M CR, 100 ␮M TR, and 90 ␮M TS
for seven days and MB for the last 24 h. Dye uptake was quantified by flow cytometry. The percentage of positive cells corresponds to the dye
uptake of MB in panel A, whereas panel D shows dye uptake for each anionic dye. Data are mean ± SEM of three independent experiments
*p < 0.05 and **p < 0.05 determined by Student’s t test. Images captured with 10× objective.

According to our docking data, dye conformers bind
at the interface region between the ␤-sheets in all cases.
The contact residues (lower than 4.5 Å) in the binding
site for conformers are indicated in Fig. 10 were also
protein residues that form hydrogen bonds are high-
lighted. Regarding the dye molecules, one and/or two
binding modes were observed: in the first mode, dye
molecules were parallel to the ␤-sheet extension direc-
tion and the second binding mode is parallel to the
␤-strand direction.
Theoretical determinations
The
Gp of the tau PHF core protofibril-dye com-
plex was calculated via
Gsolv and
Gcoul . The
solvation contribution was calculated using a continu-
ous solvent approximation and the numerical solution
to the nonlinear Poisson–Boltzmann equation imple-
mented in the program APBS and the Coulombic
contribution using a program Coulomb within APBS
accessory [21]. The
Gnp was estimated by the change
in solvent-accessible surface area (
ASA) upon bind-
ing with the program VMD. The results are presented
in Table 1 for each system. We present the binding
energies (
Gcalc ) for the tau PHF core protofibril with
the dyes. We also show hydrogen bonds established in
the binding site with a 4.5 Å cutoff. Energy values indi-
cate that the tau-CR interaction is more favorable than
that TR and TS. The binding is driven by electrostatic
interactions, mainly the direct Coulombic contribution,
rather than hydrophobic interactions.
We determined the
Gcalc using an implicit-solvent
model in which the electrostatic contributions were
obtained by solving the Poisson-Boltzmann equa-
tion numerically and the hydrophobic effects were
accounted by using a surface area dependent nonpolar
term [21, 25, 26].
328 K.I. Lira-De León et al. / Tau Aggregation Induced by Dyes
Fig. 6. Tau aggregates are induced by anionic dyes in SH-SY5Y
cells but destabilized by treatment with MB. A) Membrane filtra-
tion assay of SH-SY5Y cells treated with 100 ␮M CR, 100 ␮M
TR, and 90 ␮M TS. B) Tau immunostaining (pro-aggregative con-
ditions). C) Membrane filtration assay with the same concentration
of dyes and co-treated with 40 ␮M of MB. D) Tau immunostaining
(anti-aggregative conditions). Aggregates were visualized using Tau
DC25 antibody and could only be observed in the absence of MB.
Tau PHF core protofibril: CR complexes
The tau PHF core protofibril-dye interaction was
analyzed by docking. We selected two complexes
Fig. 7. Electron microscopy analysis of SH-SY5Y cells treated with anionic dyes and MB. A) Positive control of filaments formed with TS
without MB treatment. B) Control cells treated only with MB. SH-SY5Y cells treated with 100 ␮M CR (C), 100 ␮M TR (D), and 90 ␮M TS
(E), with 40 ␮M MB for 24 h. Scale bar = 100 nm.
with the best interaction energy interaction for
each condition. CR dye has in its structure two
important SO− 3 moieties which help to stabilize
the complex by electrostatic interactions mainly
with charged residues such as lysine. In complex
1 of the tau PHF core protofibril-CR complex, the
residues involved are Ser341, Lys311, and Glu338.
In complex 2, the residues involved are Ser341,
Gln307, Lys311, Lys340, and Ser341. Complex 2
has a
Gcalc = −467.8 kJ mol−1 , more favorable
compared to complex 1
Gcalc = −461.3 kJ mol−1 .
This difference in the energy values may be explained
by the interaction H-bond of the SO− 3 moiety of CR
that interacts with Lys311 and Lys340 of tau and
another SO− 3 moiety that can interact with Lys311
present in complex 2, while complex 1 shows that a
SO− 3 moiety interacts only with Lys311 and this is
reflected in the energy values (Table 1). At a distance
of 4.5 Å both CR complexes exhibit 5–6 hydrogen
bonds with the tau protofibril (Fig. 10A).
Tau PHF core protofibril: TR complexes
TR also has two important SO− 3 moieties in its struc-
ture. In complex 1, the residues involved in binding are:
Ser341, Val306, Gln307, Lys311, and Ser341, whereas
in complex 2 they are Gln307, Tyr310, Gln307, and
Ser341. In complex 1, the calculated energy was more
favorable (
Gcalc = −359.7 kJ mol−1 ) compared to
 K.I. Lira-De León et al. / Tau Aggregation Induced by Dyes 329

A B
Psi General Psi Glycine
180 180
150 150
120 120
90 90
60 60
30 30
0 0
-30 -30
-60 -60
-90 -90
-120 -120
-150 -150
-180 -180
-150-120 -90 -60 -30 0 30 60 90 120 150 180 -150-120 -90 -60 -30 0 30 60 90 120 150 180
Phi Phi
Psi Proline Psi Pre-Proline
180 180
150 150
120 120
90 90
60 60
30 30
0 0
-30 -30
-60 -60
-90 -90
-120 -120
-150 -150
-180 -180
-150-120 -90 -60 -30 0 30 60 90 120 150 180 -150-120 -90 -60 -30 0 30 60 90 120 150 180
Phi Phi

Fig. 8. Modeling of tau PHF core. A) Final model for tau PHF core (model 1). Final structure multiple-threading alignments by LOMETS
(http://zhanglab.ccmb.med.umich.edu/LOMETS/ ) and iterative TASSER assembly simulations using I-TASSER server. B) Ramachandran plot
for tau PHF core shows the phi-psi torsion angles for all residues in the structure. Residues are shown as black dots. Allowed regions (orange)
and core regions (green) are indicated.

Fig. 9. Modeling of tau PHF core protofibrils. Two representations of tau PHF core protofibril. A) Along of parallel ␤-strand forming each
␤-sheet. B) Along protofibril axis.

complex 2 (
Gcalc = −209.7 kJ mol−1 ). In the case
of complex 1 only one SO− 3 moiety of TR interacts
with Lys311 and this is also reflected in the calculated
energy (Table 1). At a distance of 4.5 Å both com-
plexes of TR exhibit 5–6 hydrogen bonds with the tau
protofibril (Fig. 10B).
Tau PHF core protofibril: TS complexes
The anionic dye TS has only one SO− 3 moiety.
In complex 1, residues involved in the binding are
Gln307 and Gln336. For complex 2, residues Gln336
and Lys340 are involved in the binding the dye. In
330 K.I. Lira-De León et al. / Tau Aggregation Induced by Dyes

Fig. 10. Representative binding poses of dyes to tau PHF core protofibril interface after docking and energy minimization procedure. A) CR;
B) TR; C) TS; D) MB. The backbones of each ␤-strand are shown in tubes.

Table 1
Calculated energies (
Gcalc ) of anionic and cationic dyes binding to tau via
Gsolv ,
Gcoul , and
Gnp which were determined by theoretical
calculations
Complex
Gsolv (kJ/mol)
Gcoul (kJ/mol)
Gnp (kJ/mol)
Gcalc a (kJ/mol) Hydrogenb Bonds 4.5 Å
Tau-CR Complex 1 104.3 −534.0 −31.7 −461.3 5
Complex 2 96.6 −536.5 −27.9 −467.8 6
Tau-TR Complex 1 −3.3 −333.5 −22.9 −359.7 6
Complex 2 83.8 −271.1 −22.4 −209.7 5
Tau-TS Complex 1 83.6 −139.5 −24.8 −80.7 5
Complex 2 110.4 −105 −24.6 −19.2 2
Tau-MB Complex 1 37.5 12.8 −16.6 33.7 -
Complex 2 46.3 11.6 −15.3 42.5 -
a Theenergy calculated (
Gcalc ) is given by Ec (6).
b Thenumber of hydrogen bonds between dye and tau in the binding site was determined by the program MOE.
− The program MOE was not able to determine any hydrogen bond interactions.
 K.I. Lira-De León et al. / Tau Aggregation Induced by Dyes
both complexes, the SO− 3 moiety does not interact with
any lysine at 4.5 Å. However, in complex 2 an oxy-
gen atom of Lys340 interacts with a nitrogen atom of
the anionic dye, but this interaction is not as strong
as the lysine when it interacts with the SO− 3 moiety.
At a distance of 4.5 Å, both TS complexes exhibit 2–5
hydrogen bonds with the tau protofibril (Fig. 10C). The
CR, TR, and TS models shown here involved interac-
tions with at least one lysine that is positively charged
creating an electrostatic environment in the binding site
that stabilizes the complex with tau.
Tau PHF core protofibril: MB complexes
Tau PHF protofibrils-MB docking does not show
any electrostatic interactions (Fig. 10D) and this was
corroborated by the values of
Gcalc (Table 1). The

Gcalc for both complexes were unfavorable for the
binding of MB with the tau PHF protofibrils compared
with CR, TR, and TS. However the hydrophobic inter-
actions of MB with tau are favorable but less stable
compared to electrostatic interactions.
DISCUSSION
This is the first report that shows, in a neuronal
environment, the ability of several anionic dyes to pro-
mote tau aggregation under physiological conditions
(without overexpression of tau). Although anionic dyes
usually serve as small-molecule ligands that bind ␤-
sheet structures [7], recent reports show their capacity
as inducers of tau aggregation in vitro [6]. Our results
are consistent with these in vitro observations with CR
being the most potent inducer followed by TS and TR
[6]. Our computational data suggest that the mecha-
nism of tau polymerization requires a conformational
change of the ␤-sheet structure [29] which can be
induced by anionic dyes. According to our confocal
images, an interaction between tau and the planar dyes
seems to be possible in the cytoplasm. However, in
contrast to Bandyopadhyay et al. [9], we observed that
all three dyes are able to induce tau fibrillization in
vitro. These differences may be due to specific charac-
teristics of the different cell line used. For instance, the
293T cell line does not express tau and only when arti-
ficially high levels of tau are ectopically expressed, tau
aggregates could be observed. Thus, the neuronal SH-
SY5Y cells is a more appropriate model for studying
the processes involved in pathologic tau aggregation.
It is noteworthy that the induction of tau aggregation in
this model was performed using only anionic dyes and
331
it did not require the overexpression of mutated or trun-
cated tau protein or hyperphosphorylation inducers [8].
Therefore, our model allows studying the mechanisms
of tau aggregation under conditions that resemble the
physiological environment of AD pathogenesis. Our
model presents another hallmark of AD such as tau
truncation in the N-terminal [30], present in CR and
TS treatments.
Importantly, the morphology of the filaments
obtained with each dye is different and, according to
our results, it is now possible to work under condi-
tions that best resemble the morphology of a certain
tauopathy. For example, in the case of AD, TS would
be the best inducer; while in Pick’s disease, TR would
be the better choice [31]. Interestingly, the treatment
with MB eliminated tau aggregates and similar results
were observed in vitro assays of aggregation of tau
truncations. Moreover, in these assays the aggrega-
tion of tau was destabilized by MB, and there was
no effect on the ability of tau protein when interacting
with microtubules, thus the function of tau is conserved
[12]. On the other hand, it is important to mention the
significant features of MB for destabilization of tau
aggregation: the cationic charge, planarity, and a high
level of aromatic conjugation. The last two features
appear to be determinants for its inhibitory activity
on tau aggregation, since non-conjugated and distorted
analogues display a dramatically reduced potency [12,
32]. Although the results obtained of in vivo models
for inhibition of tau aggregation with MB are contro-
versial, most of the results are consistent with their
inhibitory activity. Therefore MB and their deriva-
tives could be potential drugs to treat AD and other
tauopathies [10, 11, 33].
Tau has a very low content of secondary structural
elements as shown by sequence analysis [34]. It is
considered a natively unfolded protein as shown for
example by circular dichroism spectra [35] and retains
its disordered state even upon binding to microtubules
[17, 36]. In vitro, tau aggregation can be induced effi-
ciently only by incubation with several ligands, such
as DNA and RNA [37].
The longest human tau isoform hTau40 exists in the
central nervous system. It consists of 441 residues and
contains several domains, including the proline-rich
flanking regions P1 and P2 (residues 151–244) and
four repetitive domains R1–R4 (residues 244–369)
(Fig. 8) that constitute the MTBD with residues in
repeats R2–R4 possessing a high propensity to adopt
a ␤-structure [17, 38–40]. Atomic force microscopy,
Fourier transform infrared spectroscopy, circular
dichroism, and X-ray fiber diffraction experiments
332 K.I. Lira-De León et al. / Tau Aggregation Induced by Dyes
have shown that fragments of the ␤-structure in
these repeats can stably interact with each other and
with themselves and can aggregate into fibrils [35,
41, 42]. Petterson and co-workers [43] suggested
that the hexapeptide motifs 275 VQIINK280 and
306 VQIVYK311 in repeats R2 and R3, respectively, act
as mediators of intermolecular interactions between
tau monomers which were stabilized by the anionic
factor heparin. Electron paramagnetic resonance
spectra and electron microscopy measurements also
indicated that fragments in the R2 repeat of tau have
a parallel in-register arrangement of ␤-strands [44].
However, the highest propensity of ␤-structure content
in the tau protein is in repeats R2, R3, and R4 (Fig. 8).
Monte Carlo simulations and molecular dynamics
[45] of the 306 VQIVYK311 -domainin repeat R3 have
shown that the tau molecule not only promotes the
formation of stable oligomers but also is capable of
stabilizing these structures because such fragments
are characteristically rich in ␤-sheet structures.
We observed in our models that the electrostatic
interactions and the hydrogen bonds determine mainly
the interaction between the anionic dyes and tau PHF
core protofibril and to a lesser extent the hydrophobic
interactions (Table 1). CR has more interactions with
the lysine residues present in the protofibrils compared
to TR and TS. The presence of two SO− 3 moieties in
the structure of CR allows interaction with at least one
lysine residue and provides the electrostatic environ-
ment favoring the stability of the complex. For tau-TR
interaction, we observed that in one of the complexes,
only one SO− 3 moiety interacts with a lysine residue.
In the case of tau-TS interaction, this dye has only
one SO− 3 moiety that does not interact with any lysine
residue to 4.5 Å resulting in a weaker interaction. How-
ever, a lysine residue interacts with a nitrogen atom of
the TS, but this interaction is not as strong compared
to the lysine interaction with the SO− 3 moiety.
All dyes intercalate into ␤-sheets in tau PHF core-
protofibrils. Two binding modes have been observed.
In the first mode, the dye molecules are aligned along
the axis of tau PHF core protofibrils so that the struc-
ture is stabilized by electrostatic interactions between
the negatively charged sulfate groups present in the
dye and the positively charged amino acids in the
tau PHF core-protofibril. In the other mode, the dye
molecules are aligned in parallel between two ␤-
sheets of the tau PHF core-protofibrils. As in the first
mode, the interaction occurs between the sulfate group
of the dye and the lysine residues in the tau PHF
core-protofibrils. Regardless of how anionic dyes bind
to tau PHF core-protofibrils, the end result outcome
is the stabilization of the structure through electro-
static interactions between the negative and positive
charges. By contrast, MB binds to ␤-strands of tau
PHF core-protofibril in a parallel fashion that does not
allow establishing electrostatic interactions or hydro-
gen bonds. Instead, the binding is driven by weaker
nonpolar interactions (hydrophobic and Van der Waals)
(Table 1).
Although the mechanism of tau-anionic dyes com-
plex formation is still unknown, our theoretical models
indicate that the interactions leading to the binding of
anionic dyes with tau are of electrostatic nature. The
present study also demonstrates the usefulness of the-
oretical tools to provide a qualitative understanding of
the driving forces responsible for the binding of anionic
dyes to tau PHF core-protofibrils. In summary, our
computational model and our electron microscopy data
suggest MB as a potential destabilizing agent of tau
aggregates. This knowledge may be useful for finding
new therapeutic approaches for AD and other related
diseases.
ACKNOWLEDGMENTS
We thank Dr. Sergio Roberto Aguilar-Ruiz, Dr.
Honorio Torres-Aguilar, and Dra. Gabriela Pérez-
González for intellectual and technical support. We
thank Dr. Michael Schnoor for the critical reading to
the manuscript. I.N.S. acknowledges Support of the
Committee for Aid and Education in Neurochemistry
(CAEN) of category 1B: Research supplies for use
in the applicant’s home laboratory. K.I. Lira-De León
received predoctoral fellowships from CONACyT and
CINVESTAV. This work was partially supported by
CONACyT grant No. 101996.
Authors’ disclosures available online (http://www.j-
alz.com/disclosures/view.php?id=1650).
REFERENCES
[1] Robert M, Mathuranath PS (2007) Tau and tauopathies. Neu-
rology India 55, 11-16.
[2] Chirita CN, Kuret J (2004) Evidence for an intermediate in
tau filament formation. Biochemistry 43, 1704-1714.
[3] Chirita C, Necula M, Kuret J (2004) Ligand-dependent inhi-
bition and reversal of tau filament formation. Biochemistry
43, 2879-2887.
[4] Buee L, Bussiere T, Buee-Scherrer V, Delacourte A, Hof
PR (2000) Tau protein isoforms, phosphorylation and role
in neurodegenerative disorders. Brain Res Brain Res Rev 33,
95-130.
[5] Mandelkow E, von Bergen M, Biernat J, Mandelkow EM
(2007) Structural principles of tau and the paired helical fila-
ments of Alzheimer’s disease. Brain Pathol 17, 83-90.
 K.I. Lira-De León et al. / Tau Aggregation Induced by Dyes
[6] Chirita CN, Congdon EE, Yin H, Kuret J (2005) Triggers of
full-length tau aggregation: A role for partially folded inter-
mediates. Biochemistry 44, 5862-5872.
[7] Kuret J, Congdon EE, Li G, Yin H, Yu X, Zhong, Q
(2005) Evaluating triggers and enhancers of tau fibrillization.
Microsc Res Techn 67, 141-155.
[8] Kuret J, Chirita CN, Congdon EE, Kannanayakal T, Li G, Nec-
ula M, Yin H, Zhong Q (2005) Pathways of tau fibrillization.
Biochim Biophys Acta 1739, 167-178.
[9] Bandyopadhyay B, Li G, Yin H, Kuret J (2007) Tau aggrega-
tion and toxicity in a cell culture model of tauopathy. J Biol
Chem 282, 16454-16464.
[10] Medina DX, Caccamo A, Oddo S (2011) Methylene blue
reduces abeta levels and rescues early cognitive deficit by
increasing proteasome activity. Brain Pathol 21, 140-149.
[11] Gura T (2008) Hope in Alzheimer’s fight emerges from unex-
pected places. Nat Med 14, 894.
[12] Oz M, Lorke DE, Petroianu GA (2009) Methylene blue and
Alzheimer’s disease. Biochem Pharmacol 78, 927-932.
[13] Zilka N, Filipcik P, Koson P, Fialova L, Skrabana R, Zilkova
M, Rolkova G, Kontsekova E, Novak M (2006) Truncated tau
from sporadic Alzheimer’s disease suffices to drive neurofib-
rillary degeneration in vivo. FEBS Lett 580, 3582-3588.
[14] Chang E, Kuret J (2008) Detection and quantification of tau
aggregation using a membrane filter assay. Anal Biochem 373,
330-336.
[15] Askanas V, Engel WK, Bilak M, Alvarez RB, Selkoe DJ
(1994) Twisted tubulofilaments of inclusion body myositis
muscle resemble paired helical filaments of Alzheimer brain
and contain hyperphosphorylated tau. Am J Pathol 144, 177-
187.
[16] Chemical Computing Group, Montreal, Canada, http://www.
chemcomp.com
[17] Mukrasch MD, Biernat J, von Bergen M, Griesinger C, Man-
delkow E, Zweckstetter M (2005) Sites of tau important
for aggregation populate {beta}-structure and bind to micro-
tubules and polyanions. J Biol Chem 280, 24978-24986.
[18] Mukrasch MD, Bibow S, Korukottu J, Jeganathan S, Bier-
nat J, Griesinger C, Mandelkow E, Zweckstetter M (2009)
Structural polymorphism of 441-residue tau at single residue
resolution. PLoS Biol 7, e34.
[19] Luhrs T, Ritter C, Adrian M, Riek-Loher D, Bohrmann
B, Dobeli H, Schubert D, Riek R (2005) 3D structure of
Alzheimer’s amyloid-beta (1-42) fibrils. Proc Natl Acad Sci
U S A 102, 17342-17347.
[20] Landau M, Sawaya MR, Faull KF, Laganowsky A, Jiang L,
Sievers SA, Liu J, Barrio JR, Eisenberg D (2011) Towards a
pharmacophore for amyloid. PLoS Biol 9, e1001080.
[21] Baker NA, Sept D, Joseph S, Holst MJ, McCammon JA
(2001) Electrostatics of nanosystems: Application to micro-
tubules and the ribosome. Proc Natl Acad Sci U S A 98,
10037-10041.
[22] Dolinsky TJ, Nielsen JE, McCammon JA, Baker NA
(2004) PDB2PQR: An automated pipeline for the setup of
Poisson–Boltzmann electrostatics calculations. Nucleic Acids
Res 32, W665-W667.
[23] MacKerell AD, Bashford D, Dunbrack RL, Evanseck JD,
Field MJ, Fischer S, Gao J, Guo H, Ha S, Joseph-McCarthy D,
Kuchnir L, Kuczera K, Lau FTK, Mattos C, Michnick S, Ngo
T, Nguyen DT, Prodhom B, Reiher WE, Roux B, Schlenkrich
M, Smith JC, Stote R, Straub J, Watanabe M, Wiórkiewicz-
Kuczera J, Yin D, Karplus M (1998) All-Atom empirical
potential for molecular modeling and dynamics studies of
proteins. J Phys Chem B 102, 3586-3516.
333
[24] Li H, Robertson AD, Jensen JH (2005) Very fast empirical
prediction and rationalization of protein pKa values. Proteins
61, 704-721.
[25] Gorham RD Jr., Kieslich CA, Morikis D (2011) Electrostatic
clustering and free energy calculations provide a foundation
for protein design and optimization. Ann Biomed Eng 39,
1252-1263.
[26] Friedman RA, Honig B (1995) A free energy analysis of
nucleic acid base stacking in aqueous solution. Biophys J 69,
1528-1535.
[27] von Bergen M, Barghorn S, Li L, Marx A, Biernat J, Man-
delkow EM, Mandelkow E (2001) Mutations of tau protein in
frontotemporal dementia promote aggregation of paired heli-
cal filaments by enhancing local beta-structure. J Biol Chem
276, 48165-48174.
[28] Sawaya MR, Sambashivan S, Nelson R, Ivanova MI, Sievers
SA, Apostol MI, Thompson MJ, Balbirnie M, Wiltzius JJ,
McFarlane HT, Madsen AO, Riekel C, Eisenberg D (2007)
Atomic structures of amyloid cross-beta spines reveal varied
steric zippers. Nature 447, 453-457.
[29] von Bergen M, Barghorn S, Biernat J, Mandelkow EM, Man-
delkow E (2005) Tau aggregation is driven by a transition
from random coil to beta sheet structure. Biochim Biophys
Acta 1739, 158-166.
[30] Binder LI, Guillozet-Bongaarts AL, Garcia-Sierra F, Berry
RW (2005) Tau, tangles, and Alzheimer’s disease. Biochim
Biophys Acta 1739, 216-223.
[31] Arima K (2006) Ultrastructural characteristics of tau filaments
in tauopathies: Immuno-electron microscopic demonstration
of tau filaments in tauopathies. Neuropathology 26, 475-483.
[32] Bulic B, Pickhardt M, Mandelkow EM, Mandelkow E (2010)
Tau protein and tau aggregation inhibitors. Neuropharmacol-
ogy 59, 276-289.
[33] Wischik CM, Edwards PC, Lai RY, Roth M, Harrington
CR (1996) Selective inhibition of Alzheimer disease-like tau
aggregation by phenothiazines. Proc Natl Acad Sci U S A 93,
11213-11218.
[34] Barghorn S, Mandelkow E (2002) Toward a unified scheme
for the aggregation of tau into Alzheimer paired helical fila-
ments. Biochemistry 41, 14885-14896.
[35] von Bergen M, Friedhoff P, Biernat J, Heberle J, Man-
delkow EM, Mandelkow E (2000) Assembly of tau protein
into Alzheimer paired helical filaments depends on a local
sequence motif ((306)VQIVYK (311)) forming beta struc-
ture. Proc Natl Acad Sci U S A 97, 5129-5134.
[36] Santarella RA, Skiniotis G, Goldie KN, Tittmann P, Gross H,
Mandelkow EM, Mandelkow E, Hoenger A (2004) Surface-
decoration of microtubules by human tau. J Mol Biol 339,
539-553.
[37] Meraz-Rios MA, Lira-De Leon KI, Campos-Pena V, De
Anda-Hernandez MA, Mena-Lopez R (2010) Tau oligomers
and aggregation in Alzheimer’s disease. J Neurochem 112,
1353-1367.
[38] Margittai M, Langen R (2004) Template-assisted filament
growth by parallel stacking of tau. Proc Natl Acad Sci U S
A 101, 10278-10283.
[39] Mukrasch MD, von Bergen M, Biernat J, Fischer D,
Griesinger C, Mandelkow E, Zweckstetter M (2007) The
“jaws” of the tau-microtubule interaction. J Biol Chem 282,
12230-12239.
[40] Eliezer D, Barre P, Kobaslija M, Chan D, Li X, Heend L
(2005) Residual structure in the repeat domain of tau: Echoes
of microtubule binding and paired helical filament formation.
Biochemistry 44, 1026-1036.
334 K.I. Lira-De León et al. / Tau Aggregation Induced by Dyes
[41] Goux WJ, Kopplin L, Nguyen AD, Leak K, Rutkofsky M,
Shanmuganandam VD, Sharma D, Inouye H, Kirschner DA
(2004) The formation of straight and twisted filaments from
short tau peptides. J Biol Chem 279, 26868-26875.
[42] Barrantes A, Sotres J, Hernando-Perez M, Benitez MJ, de
Pablo PJ, Baro AM, Avila J, Jimenez JS (2009) Tau aggre-
gation followed by atomic force microscopy and surface
plasmon resonance, and single molecule tau-tau interaction
probed by atomic force spectroscopy. J Alzheimers Dis 18,
141-151.
[43] Peterson DW, Zhou H, Dahlquist FW, Lew J (2008) A soluble
oligomer of tau associated with fiber formation analyzed by
NMR. Biochemistry 47, 7393-7404.
[44] Margittai M, Langen R (2006) Side chain-dependent stacking
modulates tau filament structure. J Biol Chem 281, 37820-
37827.
[45] Li DW, Mohanty S, Irback A, Huo S (2008) Formation and
growth of oligomers: A Monte Carlo study of an amyloid tau
fragment. PLoS Comput Biol 4, e1000238.