Fluorescent BODIPY-based Zn(II) complex as a molecular probe for selective detection of neurofibrillary tangles in the brains of Alzheimer's disease patients.

Article abstract of DOI:10.1021/ja9008369

We have developed a new fluorescent binuclear Zn(II) complex for the detection of neurofibrillary tangles (NFTs) of hyperphosphorylated tau proteins, a representative hallmark of Alzheimer's disease (AD). The probe 1 incorporates a fluorescent BODIPY unit

Full text of DOI:10.1021/ja9008369

Published on Web 04/15/2009
Fluorescent BODIPY-Based Zn(II) Complex as a Molecular
Probe for Selective Detection of Neurofibrillary Tangles in the
Brains of Alzheimer’s Disease Patients
Akio Ojida,†,‡ Takashi Sakamoto,† Masa-aki Inoue,† Sho-hei Fujishima,†
Guy Lippens,§ and Itaru Hamachi*,†,|
Department of Synthetic Chemistry and Biological Chemistry, Graduate School of Engineering,
Kyoto UniVersity, Katsura campus, Kyoto 615-8510, Japan, JST, PRESTO (Life Phenomena and
Measurement Analysis), Sanbancho, Chiyodaku, Tokyo 102-0075, Japan, CNRS-UniVersite´ de
Lille 1, UMR 8576, USTL, AVenue de MandeleieV, 59655 VilleneuVe d’Ascq Cedex, France, and
JST, CREST (Creation of Next-generation Nanosystems through Process Integration),
Sanbancho, Chiyodaku, Tokyo 102-0075, Japan
Received February 4, 2009; E-mail: ihamachi@sbchem.kyoto-u.ac.jp
Abstract: We have developed a new fluorescent binuclear Zn(II) complex for the detection of neurofibrillary
tangles (NFTs) of hyperphosphorylated tau proteins, a representative hallmark of Alzheimer’s disease (AD).
The probe 1 incorporates a fluorescent BODIPY unit and two Zn(II)-2,2′-dipicolylamine (Dpa) complexes
as a binding site for phosphorylated amino acid residues. Using fluorescence titration to evaluate the binding
and sensing properties of 1 toward several phosphorylated peptide segments derived from hyperphos-
phorylated tau protein, we found that 1 binds preferentially to peptides presenting phosphorylated groups
at the i and i+4 positions with dissociation constants (Kd) in the micromolar range. Fluorescence titration
with an artificially prepared aggregate of the phosphorylated tau protein (p-Tau) revealed that 1 binds strongly
to p-Tau (EC50 ) 9 nM). In contrast, the interactions of 1 were weaker toward artificially prepared aggregates
of the nonphosphorylated tau protein (n-Tau; EC50 ) 80 nM) and Aꢀ1-42 fibrils (EC50 ) 650 nM). Histological
imaging of a hippocampus tissue section obtained from an AD patient revealed that 1 fluorescently visualizes
deposits of NFTs and clearly discriminates between NFTs and the amyloid plaques assembled from
amyloid-ꢀ peptides, confirming our strategy toward the rational design of a molecular probe for the selective
fluorescence detection of NFTs in brain tissue sections.
Introduction
at more than 30 sites and, as a result, loses its binding ability
to microtubles.3 The liberated hyperphosphorylated tau ac-
Phosphorylation is a prevalent regulatory mechanism of many
protein functions; it plays a central role in many biological
processes.1 Protein phosphorylation is usually controlled in terms
of the number of phosphate units and their position through
the individual action of protein kinases and phosphatases. An
overall imbalance of their activities can cause aberrant hyper-
phosphorylation of proteins, which sometimes leads to serious
disruption of cellular functions.2 A representative example is
the hyperphosphorylation of tau proteins. Tau is a microtubule-
associated protein that is present in high abundance in neuronal
cells. The binding of tau to microtubules is primarily regulated
by serine/threonine-directed phosphorylations. Under pathologi-
cal conditions, however, tau is abnormally hyperphosphorylated
cumulates in brain tissue as insoluble filamentous aggregates,
so-called neurofibrillary tangles (NFTs).4 It is widely recognized
that insoluble deposits of NFTs and the senile plaques (SPs)
assembled from ꢀ-amyloid peptide (Aꢀ) are the representative
pathological hallmarks of many neurodegenerative disorders,
including Alzheimer’s disease (AD).5 In the past two decades,
many studies have been devoted to understanding of the
deleterious effects of NFTs and SPs in the progress of dementia;
nevertheless, the mechanistic linkages between these two fibrillar
formations and their functional relationships in various neuro-
degenerative diseases remain controversial.6 The development
of a functional molecular probe that enables the individual
† Kyoto University.
(3) (a) Hanger, D. P.; Byers, H. L.; Wray, S.; Leung, K. Y.; Saxton, M. J.;
Seereeram, A.; Reynolds, C. H.; Ward, M. A.; Anderton, B. H. J. Biol.
Chem. 2007, 282, 23645–23654. (b) Lindwall, G.; Cole, R. D. J. Biol.
Chem. 1984, 259, 5301–5305.
‡ JST, PRESTO.
§ CNRS-Universite´ de Lille 1.
| JST, CREST.
(1) (a) Hunter, T. Protein Phosphorylation; Academic Press: New York,
1998. (b) Pawson, T.; Scott, J. D. Trends Biochem. Sci. 2005, 30,
286–290.
(4) Lee, V. M.; Balin, B. J.; Otvos, L., Jr.; Trojanowski, J. Q. Science
1991, 251, 675–678.
(5) Goedert, M.; Spillantini, M. G. Science 2001, 314, 777–781.
(6) (a) King, M. E.; Kan, H. M.; Baas, P. W.; Erisir, A.; Glabe, C. G.;
Bloom, G. S. J. Cell. Biol. 2006, 175, 541–546. (b) Leschik, J.; Welzel,
A.; Weissmann, C.; Eckert, A.; Brandt, R. J. Neurochem. 2007, 101,
1303–1315. (c) Rapoport, M.; Dawson, H. N.; Binder, L. I.; Vitek,
M. P.; Ferreira, A. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 6364–
6369.
(2) (a) Go¨tz, J.; Ittner, L. M. Nat. ReV. Neurosci. 2008, 9, 532–544. (b)
Chakraborti, S.; Das, S.; Kar, P.; Ghosh, B.; Samanta, K.; Kolley, S.;
Ghosh, S.; Roy, S.; Chakraborti, T. Mol. Cell. Biochem. 2007, 298,
1–40. (c) Veeramani, S.; Yuan, T. C.; Chen, S. J.; Lin, F. F.; Petersen,
J. E.; Shaheduzzaman, S.; Srivastava, S.; MacDonald, R. G.; Lin, M. F.
Endocr. Relat. Cancer 2005, 12, 805–822.
9
10.1021/ja9008369 CCC: $40.75  2009 American Chemical Society
J. AM. CHEM. SOC. 2009, 131, 6543–6548 6543

A R T I C L E S
Ojida et al.
Preparation of in Vitro Aggregates. Aggregates of the p-Tau
and n-Tau proteins were prepared according to the reported
method.11 Briefly, the purified p-Tau or n-Tau protein (8 µM) was
incubated with heparin (1.6 µM) in 10 mM HEPES buffer (pH
7.6) containing 40 mM NaCl, 0.1 mM EDTA, and 5 mM DTT at
37 °C for 1-20 days. The formation of the aggregates was
confirmed using transmitted electron microscopy (TEM) and
fluorescence titration with thioflavin-T. The Aꢀ1-42 fibril was
prepared according to a reported method.12 Briefly, an aqueous
ammonium hydroxide solution (0.02%) of Aꢀ1-42 peptide (250 µM,
Wako Pure Chemical) was diluted 10-fold with HBS and left to
stand for 3 days at 37 °C.
Fluorescence Titration of in Vitro Aggregates With Probe
1. A suspension of p-Tau (8 µM) or Aꢀ1-42 aggregate (25 µM)
was sequentially diluted (1:25 to 1:6500 for p-Tau; 1:2.5 to 1:320
for Aꢀ) with HBS buffer [containing 10% (v/v) DMSO and 10
µM Zn(NO3)2], and then each diluted solution was mixed with a
solution of 1 (100 nM) in HBS buffer. After incubation for 2.5 h
at 37 °C, the fluorescence intensity (545 nm) of each solution was
measured using a fluorescence microplate reader (Infinite M200,
TECAN) with excitation at 490 nm.
detection of NFTs and SPs would benefit approaches based on
fluorescence imaging or positron emission tomography (PET)
to allow a more precise understanding of the pathophysiology
of neurodegenerative diseases. For this purpose, several small
molecular probes have been developed,7 but none of them has
been able to discriminate between NFTs and SPs with high
specificity.8
In this Article, we report the rational design, function, and
application of the binuclear Zn(II) complex 1 as a fluorescent
probe for the selective detection of NFTs in brain tissue. The
strong binding of 1 toward aggregated hyperphosphorylated tau
proteins enables a simple staining procedure to be used for the
selective fluorescence imaging of NFTs. Notably, a sharp
discrimination of NFTs from SPs results from the phospho-
selective binding properties of 1, based on multivalent
metal-ligand interactions with multiple phosphorylated sites
of the tau protein.
Experimental Section
Evaluation of the EC50 of 1 for the in Vitro Aggregates. A
suspension of the aggregate (1 µg/mL) of p-Tau, n-Tau, or Aꢀ1-42
was incubated with 0.3 nM-3.3 µM of 1 in HBS [containing 10%
(v/v) DMSO and 100 µM Zn(NO3)2] at 37 °C for 10 min. The
fluorescence intensity (545 nm) of each solution was measured using
a microplate reader (Infinite M200, TECAN) with excitation at 490
nm. The EC50 values (defined as the effective concentration of 1
required to achieve 50% of the maximal fluorescence change) were
calculated using IgorPro software (WaveMetrics, Inc.).
Fluorescence Staining of in Vitro Aggregates. A suspension
of p-Tau, n-Tau, or Aꢀ1-42 aggregate was diluted 4-fold with an
HBS buffer solution of 1 (10 µM) and thioflavin T (10 µM), and
then the mixture was incubated for 10 min at room temperature.
For the control experiment, sodium pyrophosphate (PPi, 100 µM)
was added to the solution. After centrifugation (15 000 rpm, 15
min, 4 °C), the supernatant was discarded, and the collected
aggregate was washed twice with 0.5 mM aqueous Zn(NO3)2. The
aggregate was resuspended in 0.5 mM aqueous Zn(NO3)2 (10 µL),
and then a portion of this suspension (0.5 µL) was placed onto a
glass slip (22 × 24 mm; thickness: 0.12-0.17 mm) and air-dried.
The fluorescence images were obtained through confocal laser
scanning microscopy (FV-1000, Olympus) using an optimal excita-
tion laser and detection wavelength window for each fluorescent
dye (probe 1, λex ) 488 nm, λem ) 560-580 nm; thioflavin T, λex
) 458 nm, λem ) 470-490 nm).
Fluorescence Immunohistochemical Staining of Brain
Tissue Sections. Paraffin-embedded human hippocampus tissue
sections from an Alzheimer’s disease patient and a healthy adult
were purchased from BioChain Institute (CA). The tissue sections
were deparaffinized using a standard procedure employing xylene
and ethanol. To remove the fluorescent deposits of lipofuscin, the
deparaffinized tissue section was incubated in 0.25 wt % KMnO4/
PBS for 30 min and then washed twice with PBS. The tissue section
was incubated in PBS containing 1 wt % oxalic acid and 1 wt %
potassium pyrisulfate for 6 min and then washed with PBS. The
tissue section was treated with trypsin (0.05 wt % in PBS) for 15
min at 37 °C and then washed with PBS containing 0.2 wt % Tween
20 (PBS-Tween) to retrieve antigen activity. For the dephospho-
rylation of NFTs,13 the tissue section was treated with trypsin and
then incubated with PP2A (0.5 units, Upstate) for 24 h at 37 °C.
After blocking with 10% goat serum in PBS for 1 h at 37 °C, a
Fluorescence Titration of Tau Peptide With Probe 1. Fluo-
rescence spectra were recorded on a Perkin-Elmer LS55 spectrof-
luorophotometer. The fluorescence quantum yield of 1 was deter-
mined in neutral aqueous solution (50 mM HEPES, pH 7.2) using
rhodamine B (Sigma, SL, US) as a standard (Φ ) 0.31).9 The
titration experiments with the phosphorylated peptides were per-
formed at 25 °C using a solution of 1 (5 µM) or 3 (5 µM) in 50
mM HEPES (pH 7.2, 3 mL) in a quartz cell. The fluorescence
emission spectra (excitation wavelength; λex ) 480 nm) were
measured after addition (via a micro syringe) of a freshly prepared
aqueous solution of the peptide. Fluorescence titration curves (λem
) 545 nm) were analyzed using nonlinear least-squares curve-fitting
to evaluate the dissociation constant (Kd).
Preparation of Phosphorylated Tau (p-Tau) Protein. The
N-terminus 6His-tagged full-length tau protein (n-Tau, 447 amino
acids) was expressed in Tau441-pET15b-transformed Escherichia
coli BL21(DE3) and purified through a HisTag affinity column
according to the reported procedure.10 The purified n-Tau protein
(200 µg/mL, 4.4 µM) was phosphorylated through treatment with
GSK-3ꢀ (4.6 units/pmol of Tau protein, New England Biolabs) in
40 mM HEPES (pH 7.6; containing 5 mM EGTA, 3 mM MgCl2,
and 2 mM ATP) for 24 h at 30 °C. The buffer was exchanged to
HBS buffer through dialysis at 4 °C (MWCO ) 3500). The
concentration of the phosphorylated tau protein (p-Tau) was
determined using the BCA method with BSA as a standard. The
average phosphorylation number of p-Tau was determined through
in-gel fluorescence analysis using ProQ diamond (Invitrogen) as a
staining reagent; ovalbumin (2 mol of phosphate/mol of protein)
was used as a standard phosphorylated protein.
(7) (a) Raymond, S. B.; Skoch, J.; Hills, I. D.; Nesterov, E. E.; Swager,
T. M.; Bacskai, B. J. Eur. J. Nucl. Med. Mol. Imaging 2008, 35, S93-
98. (b) Li, Q.; Min, J.; Ahn, Y. H.; Namm, J.; Kim, E. M.; Lui, R.;
Kim, H. Y.; Ji, Y.; Wu, H.; Wisniewski, T.; Chang, Y. T. ChemBio-
Chem 2007, 8, 1679–1687. (c) Nilsson, K. P.; Aslund, A.; Berg, I.;
Nystro¨m, S.; Konradsson, P.; Herland, A.; Ingana¨s, O.; Stabo-Eeg,
F.; Lindgren, M.; Westermark, G. T.; Lannfelt, L.; Nilsson, L. N.;
Hammarstro¨m, P. ACS Chem. Biol. 2007, 2, 553–560. (d) Shin, J.;
Lee, S. Y.; Kim, S. H.; Kim, Y. B.; Cho, S. J. Neuroimage 2008, 43,
236–244. (e) Henriksen, G.; Yousefi, B. H.; Drzezga, A.; Wester, H. J.
Eur. J. Nucl. Med. Mol. Imaging 2008, 35, S75-S81. (f) Nordberg,
A. Eur. J. Nucl. Med. Mol. Imaging 2008, 35, S46-S50. (g) Nordberg,
A. Curr. Opin. Neurol. 2007, 20, 398–402.
(8) Okamura, N.; Suemoto, T.; Furumoto, S.; Suzuki, M.; Shimadzu, H.;
Akatsu, H.; Yamamoto, T.; Fujiwara, H.; Nemoto, M.; Maruyama,
M.; Arai, H.; Yanai, K.; Sawada, T.; Kudo, Y. J. Neurosci. 2005, 25,
10857–10862.
(11) Carlson, S. W.; Branden, M; Voss, K.; Sun, Q.; Rankin, C. A.;
Gamblin, T. C. Biochemistry 2007, 46, 8838–8849.
(12) Murakami, K.; Irie, K.; Morimoto, A.; Ohigashi, H.; Shindo, M.;
Nagao, M.; Shimizu, T.; Shirasawa, T. Biochem. Biophys. Res.
Commun. 2002, 294, 5–10.
(13) Liu, F.; Grundke-Iqbal, I.; Iqbal, K.; Gong, C. X. Eur. J. Neurosci.
2005, 22, 1942–1950.
(9) Magde, D.; Rogas, G. E.; Seybold, P. G. Photochem. Photobiol. 1999,
70, 737–744.
(10) Lippens, G.; Wieruszeski, J. M.; Leroy, A.; Smet, C.; Sillen, A.; Bue´e,
L.; Landrieu, I. ChemBioChem 2004, 5, 73–78.
9
6544 J. AM. CHEM. SOC. VOL. 131, NO. 18, 2009

Fluorescent BODIPY-Based Zn(II) Complex
A R T I C L E S
proteins because of its poor fluorescent properties, including a
very weak emission (Φ < 0.1) and short excitation and emission
wavelengths (<450 nm). To improve upon this performance,
in this study we designed a new probe 1 for the efficient
fluorescence detection of hyperphosphorylated tau protein. In
the molecular design of 1, we employed BODIPY as the
fluorescent unit; it displays a bright fluorescence in the visible
region (>500 nm). We connected the two Zn(II)-Dpa units
directly to the BODIPY moiety (i.e., without using alkyl linkers)
to provide the system with a rather planar and rigid conforma-
tion, which we suspected would provide stronger interactions
with the ꢀ-sheet-rich structure of the NFTs. We introduced two
triethylene glycol chains to impart the probe with sufficient
solubility in aqueous solutions. Scheme 1 outlines our synthetic
route toward 1. Briefly, the acid-catalyzed condensation between
the aldehyde 4 and 2,4-dimethylpyrrrole yielded 5, which we
treated with iodine and then complexed with boron trifluoride
to give 7. Compound 7 was converted to 9 via Suzuki coupling
with the pyridyl pinacolylborane 8; subsequent acid-catalyzed
deprotection and reductive amination with 2-aminomethylpi-
coline gave the bis-Dpa ligand 11. Finally, complexation with
2 equiv of Zn(NO3)2 yielded the probe 1, which we characterized
using 1H NMR spectroscopy and elemental analysis. Details of
the synthetic procedure of 1 are provided in the Supporting
Information. As a control fluorescence probe, we also prepared
the mononuclear Zn(II)-Dpa complex 2 using a synthetic route
similar to that for 1 (Supporting Information).
Figure 1. Structures of probes 1-3.
solution of primary mouse monoclonal antibodies AT8 (1:5000,
Innogenetics) or Tau2 (1:5000, Sigma) was poured onto the tissue
section, which was then incubated overnight at 4 °C. After being
washed with PBS-Tween on ice (5 × 2 min), the tissue section
was incubated with antimouse IgG goat IgG (Invitrogen, 1:500)
labeled with AlexaFluor 633 in 5% goat serum/PBS for 1 h at 37
°C, and then it was washed with PBS-Tween on ice (3 × 2 min).
The tissue sections were rinsed with HBS buffer and incubated
with 0.0001 wt % DAPI and 10 µM of 1 in HBS for 10 min at
room temperature. After being washed twice with ice-cooled 0.5
mM aqueous Zn(NO3)2, the tissue section was coated with aqueous
mounting media (PermaFluor, Beckman Coulter) and subjected for
fluorescence imaging through confocal laser scanning microscopy
(FV-1000, Olympus) using an optimal excitation laser and detection
wavelength window for each fluorescent dye (DAPI, λex ) 351
nm, λem ) 400-500 nm; 1, λex ) 488 nm, λem ) 500-555 nm;
AlexaFluor 633, λex ) 630 nm, λem ) 645-745 nm). Several
fluorescence images of the same tissue section were merged into
one, using image processing software (Photoshop 6.0; Adobe
Systems, Inc.).
Fluorescence Titration with Phosphorylated Tau
Fragment Peptides. The absorption and emission maxima of
probe 1 appear at 520 and 547 nm, respectively. For our
purposes, the fluorescence quantum yield of 1 was sufficiently
high (Φ ) 0.31) under neutral aqueous conditions (50 mM
HEPES, pH 7.2). Initially, we evaluated the sensing ability of
1 through fluorescence titration with a series of phosphorylated
peptides (Table 1), all of which are fragment peptides derived
from the hyper-phosphorylated human tau protein. Figure 2
displays a typical fluorescence change for the titration with the
tau(227-238)-2P peptide, which possesses phospho-threonine
and -serine residues at the (i, i+4) positions. The emission
intensity of 1 increased by up to ca. 1.5-fold, with a typical
saturation behavior, upon the gradual addition of tau(227-238)-
2P peptide. Through curve-fitting analysis of the emission
change, we estimated the dissociation constant (Kd) to be 7.5
µM. The fluorescence Job’s plot indicated that the binding
stoichiometry was 1:1 (Supporting Information). We obtained
a similar value for the dissociation constant (Kd ) 9.1 µM; n )
1.06; ∆H ) 7.13 kcal/mol; T∆S ) 13.9 kcal/mol) when using
isothermal titration calorimetry (ITC; see the Supporting
Information), confirming that the emission response of 1 directly
reflects its binding with the tau(227-238)-2P peptide. We
observed no changes in fluorescence for mixtures of 1 with
either the mono- or the nonphosphorylated tau(227-238)-1P
and tau(227-238)-0P peptide (Figure 2b). The fluorescence of
the mononuclear Zn(II) complex 2 did not change after the
addition of bis-phosphorylated tau(227-238)-2P (data not
shown). These results suggest that two-point binding, using both
Zn(II)-Dpa sites of 1, is essential for strong binding to and
fluorescence sensing of bis-phosphorylated peptides. Table 1
reveals that 1 can also fluorescently sense the bis-phosphorylated
tau(394-403)-2P and tau(204-217)-2 Pa and peptides and the
tris-phosphorylated Tau(204-216)-3P peptide, all of which
possess two-phosphorylated residues at (i, i+4) positions.
Results and Discussion
Design and Synthesis of the BODIPY-Based Fluorescent
Probe. Recently, we reported that the binuclear zinc complexes
of 2,2′-dipicolylamine (Dpa) serve as artificial receptors for
phosphorylated peptides.14,15 For example, the bipyridyl-type
Zn(II) complex 3 (Figure 1) binds to a bis-phosphorylated
peptide with high affinity (Kd ) ca. 0.1 µM) via a two-point
metal-ligand interaction between the Zn(II)-Dpa sites and the
pair of phosphate groups. The complex 3 is not applicable,
however, to the fluorescence detection of hyperphosphorylated
(14) (a) Ojida, A.; Hamachi, I. Bull. Chem. Soc. Jpn. 2006, 79, 35–46. (b)
Ojida, A.; Inoue, M.; Mito-oka, Y.; Tsutsumi, H.; Sada, K.; Hamachi,
I. J. Am. Chem. Soc. 2006, 128, 2052–2058. (c) Ojida, A.; Mito-oka,
Y.; Sada, K.; Hamachi, I. J. Am. Chem. Soc. 2004, 126, 2454–2463.
(d) Ojida, A.; Inoue, M.; Mito-Oka, Y.; Hamachi, I. J. Am. Chem.
Soc. 2003, 125, 10184–10185.
(15) Grauer, A.; Riechers, A.; Ritter, S.; Ko¨nig, B. Chem.-Eur. J. 2008,
14, 8922.
9
J. AM. CHEM. SOC. VOL. 131, NO. 18, 2009 6545

A R T I C L E S
Ojida et al.
Scheme 1. Synthesis of Probe 1
Table 1. Dissociation Constants (Kd) for the Binding of 1 to Tau Peptidesa
a Values of Kd were determined through analysis of fluorescence titration curves using nonlinear least-squares curve-fitting.b This value of Kd was not
determined because only a small change in fluorescence intensity occurred after addition of the peptide.
On the other hand, no fluorescence change was induced by
the tau(231-238)-2P and tau(204-217)-2Pb peptides, which
possess their two phosphorylated residues at (i, i+2) and (i,
i+6) positions, respectively. These results suggest that 1 binds
preferably to bis-phosphorylated (i, i+4) peptides. We ascribe
this sequence selectivity to the rigid structure of 1, which results
in rather strict distance-discrimination of the two phosphorylated
residues in the peptides.
Selective Binding to in Vitro Aggregates of Hyperphospho-
rylated Tau Protein. Next, we evaluated the binding ability of
1 toward the in vitro aggregate of the phosphorylated tau protein
(p-Tau). p-Tau was obtained through treatment of the recom-
binant full-length tau441 protein with GSK-3ꢀ (glycogen
synthase kinse-3ꢀ), a representative protein kinase associated
with the hyperphosphorylation event of tau protein.3a Quantita-
tive in-gel fluorescence analysis revealed that each protein
presented an average of 6.1 phosphorylation sites (see the
Supporting Information for details). Using a literature method,11
we performed the fibrillization of p-Tau through long-time
incubation (>24 h) with heparin as an aggregation promoter;
the morphology of the protein aggregate was confirmed through
TEM analysis (Supporting Information). Figure 3a reveals a
significant increase in the fluorescence of 1 after addition of
the p-Tau aggregate in the nanomolar concentration range,
suggesting that the binding affinity of 1 was greatly enhanced
as compared to that for the phosphorylated peptide fragments
(Table 1). From this fluorescence binding assay, we obtained a
value of 9 nM for the EC50 of 1 toward the p-Tau aggregate
(Figure 3b). We ascribe this affinity enhancement to a multi-
valent binding effect in the coordination of 1 with the multiple
phosphorylation sites of p-Tau, as well as to the hydrophobic
interactions between the planar BODIPY unit of 1 and the
ꢀ-sheet-rich structure of the fibrils. This binding affinity is
significantly stronger than that for the in vitro aggregate prepared
from nonphosphorylated tau (n-Tau; EC50 ) 80 nM), which
we obtained using a similar procedure. These results clearly
indicate that 1 exhibits phospho-selective binding behavior.
Interestingly, we obtained a value of 650 nM for the EC50 of
the peptide fibril composed of Aꢀ1-42 (Figure 3b), suggesting
that the interaction of 1 with the Aꢀ1-42 fibril is much weaker
by ca. 70-fold relative to that of the p-Tau protein aggregate.
This finding is reasonable when considering the fact that
9
6546 J. AM. CHEM. SOC. VOL. 131, NO. 18, 2009

Fluorescent BODIPY-Based Zn(II) Complex
A R T I C L E S
Figure 3. Fluorescence titration profiles of 1 with in vitro aggregates of
p-Tau, n-Tau, and Aꢀ1-42. (a) Changes in the fluorescence emission of 1
(100 nM) upon the addition of p-Tau (orange) and Aꢀ1-42 aggregate (green);
error bars represent s.d. (b) Fluorescence binding assay for the interactions
of 1 with the aggregates (1 µg/mL) of p-Tau (orange), n-Tau (blue), and
Aꢀ1-42 (green); error bars represent s.d.
Figure 2. Fluorescence titration analysis for the binding of 1 with
phosphorylated tau peptides. (a) Fluorescence spectral change of 1 (5 µM)
upon addition of the bis-phosphorylated tau peptide, tau(227-238)-2P (0,
2, 4, 6, 8, 10, 13, 16, 19, 22, 25 µM). (b) Plot of the fluorescence intensity
of 1 (λem ) 545 nm) upon addition of the bis-phosphorylated tau(227-238)-
2P (b), monophosphorylated tau(227-238)-1P ([), and nonphosphorylated
tau(227-238)-0P (0) peptides. Measurement conditions: 50 mM HEPES,
pH 7.2, 25 °C.
multivalent coordination is unlikely to exist in the binding with
the Aꢀ1-42 fibril. We used fluorescence microscopy to visualize
the interactions of 1 with these aggregates (Figure 4). The
fluorescence of 1 was clearly evident on the fibrous aggregates
of p-Tau deposited on a glass slide (Figure 4e); it coincided
well with the fluorescence image of the sample stained with
thioflavin-T, a nonselective fluorescent probe for ꢀ-sheet
structures of protein aggregates (Figure 4a).16 Interestingly, the
fluorescence image of 1 almost disappeared when we conducted
the staining process in the presence of a high concentration of
pyrophosphate anions (100 µM), a competitive binder for the
Zn(II)-Dpa unit of 1 (Figure 4f).17 In addition, neither the
nonphosphorylated n-Tau nor Aꢀ1-42 aggregates exhibited
fluorescence behavior when stained with 1 (Figure 4g and h,
respectively). Thus, these fluorescence microscopy analyses
further supported the notion of selective binding of 1 to the
p-Tau aggregate; more importantly, they demonstrated the
potential for using 1 to visualize NFTs in brain tissue with
sufficient discrimination from SPs comprising aggregates of Aꢀ.
Histological Staining of NFTs in AD Brain Tissue. Next, we
used 1 as a fluorescent staining agent for histological imaging
of a hippocampus tissue section from the brain of an AD patient
(Figure 5). Short-term incubation (10 min) of the tissue section
Figure 4. Fluorescence microscopy analyses of the in vitro aggregates of
p-Tau, n-Tau, and Aꢀ1-42 stained with (a-d) thioflavin T and (e-h) 1. In
(b) and (f), the staining of p-Tau was conducted in the presence of sodium
pyrophosphate (100 µM). Scale bars: 10 µm.
with 1 (10 µM) and subsequent washing with an aqueous
solution of Zn(NO3)2 (0.5 mM) afforded a stained sample that
exhibited many bright dots under fluorescence microscopy
analysis (Figure 5a and d). Gratifyingly, these dots coincided
with those in the immuno-staining image of NFTs obtained using
AT8, a monoclonal antibody for p-Tau (Figure 5b and e). The
fluorescence image of 1 also overlapped well with the tissue
section image obtained through staining with Tau2, a phospho-
rylation-independent monoclonal antibody for the tau protein
(Figure 5g-i). In contrast, fluorescent dots were scarce in the
samples of the normal hippocampus tissue section that had been
stained with 1 and AT8 (Figure 5p-r). These results strongly
suggested that 1 can be used to fluorescently visualize NFTs
deposited in the brain tissues of AD patients. The fluorescence
image of 1 almost disappeared when we pretreated the brain
tissue section with protein phosphatase 2A (PP2A),13 although
the deposited NFTs could still be detected after corresponding
(16) Leliveld, S. R.; Korth, C. Neurosci. Res. 2007, 85, 2285–2297.
(17) Ojida, A.; Takashima, I.; Kohira, T.; Nonaka, H.; Hamachi, I. J. Am.
Chem. Soc. 2008, 130, 12095–12101.
9
J. AM. CHEM. SOC. VOL. 131, NO. 18, 2009 6547

A R T I C L E S
Ojida et al.
detection of NFTs, in good agreement with the results of our
titration and staining experiments using the fibrous aggregates
in vitro (Figures 3 and 4). Our most significant finding is that
1 can be used to visualize NFTs, but not SPs composed of Aꢀ,
in the brain tissue section of an AD patient. Figure 5j-l reveals
that the deposit pattern of the NFTs was distinct from that of
the SPs; the NFTs detected by 1 are randomly distributed as
small dots, whereas the SPs detected through immuno-staining
with anti-Aꢀ antibody exist as larger (>50 µm) oval-shaped
aggregates. Overall, these results indicate the applicability of
using 1 for the selective detection of NFTs in brain tissue
sections, with sharp discrimination from SPs; indeed, they
demonstrate the validity of the molecular design of 1 as an NFT-
specific probe that operates based on multivalent coordination.
Conclusion
We have prepared a new BODIPY-type zinc complex 1 that
functions as a fluorescent probe that allows the imaging of NFTs,
materials representative of many neurodegenerative disorders
(including AD), in brain tissue samples. We exploited the
phospho-selective binding properties of 1 to realize the selective
detection of NFTs through the coordination of 1 with the
multiple phosphate groups of the hyperphosphorylated tau
protein. The probe 1 is the first example of a rationally designed
small molecule that displays a high detection selectivity for
NFTs over SPs; such selectivity is not readily achieved with
other small molecular probes, which have been designed mainly
to recognize the hydrophobic clefts of the ꢀ-sheet structure that
commonly appears in these protein aggregates. From a histo-
logical staining experiment using brain tissue from an AD
patient, 1 differentiated between the deposit patterns and
localization sites of the NFTs and SPs. Thus, 1 appears to have
great applicability for use in the precise and rapid detection of
NFTs without the need for time-consuming immuno-staining
with expensive antibodies. We believe that 1 will be an excellent
tool for the precise pathological understanding and diagnosis
of tau protein-related neurodegenerative disorders, including AD.
Further applications of 1 and related derivatives for in vivo
imaging studies are currently in progress.
Figure 5. Fluorescence immunohistochemical analyses of hippocampus
tissue sections from an AD patient. (a-c) Images of the tissue section triple-
stained with 1, AT8 (anti-p-Tau monoclonal antibody), and DAPI: (a)
staining with of 1 (green); (b) staining with AT8 (red); (c) superimposition
of images (a) and (b) and that obtained through staining with DAPI (blue).
(d-f) Magnified images of (a-c). (g-i) Images of the tissue section triple-
stained with 1, Tau2 (phosphorylation-independent monoclonal antibody
for tau), and DAPI: (g) staining with 1 (green); (h) staining with Tau2 (red);
(i) superimposition of images (g) and (h) and that obtained through staining
with DAPI (blue). (j-l) Images of the tissue section triple-stained with 1,
Aꢀ42 (anti-Aꢀ1-42 monoclonal antibody), and DAPI: (j) staining with 1
(green); (k) staining with Aꢀ42 (red); (l) superimposition of images (j) and
(k) and that obtained through staining with DAPI (blue). (m-o) Images of
the tissue section double-stained with 1 and Tau2 (phosphorylation-
independent monoclonal antibody for tau) after dephosphorylation with
PP2A: (m) staining with 1 (green); (n) staining with Tau2 (red); (o)
superimposition of images (m) and (n). (p-r) Images of hippocampus tissue
section from healthy adult, triple-stained with 1, AT8, and DAPI: (p) staining
with 1 (green); (q) staining with AT8 (red); (r) superimposition of images
(p) and (q) and that obtained through staining with DAPI (blue). Scale bars:
50 µm (a-c, g-r); 10 µm (d-f).
Acknowledgment. This study was supported by the Innovative
Techno-Hub for Integrated Medical Bioimaging Project of the
Special Coordination Funds for Promoting Science and Technology,
from the Ministry of Education, Culture, Sports, Science, and
Technology (MEXT), Japan, and by the Sumitomo Foundation,
Japan.
Supporting Information Available: Western blot analysis of
synthetic tau protein; TEM analysis of p-Tau and n-Tau fibrils;
procedures for the syntheses of 1 and 2. This material is
available free of charge via the Internet at http://pubs.acs.org.
treatment with the immuno-staining Tau2 antibody (Figure
5m-o). This result clearly indicates that coordination of 1 with
the phosphate groups of p-Tau is crucial for the effective
JA9008369
9
6548 J. AM. CHEM. SOC. VOL. 131, NO. 18, 2009