Tuning reactivity of diphenylpropynone derivatives with metal-associated amyloid-β species via structural modifications

Article abstract of DOI:10.1021/ic400851w

A diphenylpropynone derivative, DPP2, has been recently demonstrated to target metal-associated amyloid-β (metal-Aβ) species implicated in Alzheimer's disease (AD). DPP2 was shown to interact with metal-Aβ species and subsequently control Aβ aggregation (reactivity) in vitro; however, its cytotoxicity has limited further biological applications. In order to improve reactivity toward Aβ species and lower cytotoxicity, along with gaining an understanding of a structure-reactivity-cytotoxicity relationship, we designed, prepared, and characterized a series of small molecules (C1/C2, P1/P2, and PA1/PA2) as structurally modified DPP2 analogues. A similar metal binding site to that of DPP2 was contained in these compounds while their structures were varied to afford different interactions and reactivities with metal ions, Aβ species, and metal-Aβ species. Distinct reactivities of our chemical family toward in vitro Aβ aggregation in the absence and presence of metal ions were observed. Among our chemical series, the compound (C2) with a relatively rigid backbone and a dimethylamino group was observed to noticeably regulate both metal-free and metal-mediated Aβ aggregation to different extents. Using our compounds, cell viability was significantly improved, compared to that with DPP2. Lastly, modifications on the DPP framework maintained the structural properties for potential blood-brain barrier (BBB) permeability. Overall, our studies demonstrated that structural variations adjacent to the metal binding site of DPP2 could govern different metal binding properties, interactions with Aβ and metal-Aβ species, reactivity toward metal-free and metal-induced Aβ aggregation, and cytotoxicity of the compounds, establishing a structure-reactivity-cytotoxicity relationship. This information could help gain insight into structural optimization for developing nontoxic chemical reagents toward targeting metal-Aβ species and modulating their reactivity in biological systems.

Full text of DOI:10.1021/ic400851w

Article
pubs.acs.org/IC
Tuning Reactivity of Diphenylpropynone Derivatives with
Metal-Associated Amyloid‑β Species via Structural Modifications
Yuzhong Liu,†,‡,§ Akiko Kochi,†,§ Amit S. Pithadia,† Sanghyun Lee,‡,⊥ Younwoo Nam,‡,∇
Michael W. Beck,† Xiaoming He,‡,○ Dongkuk Lee,∇ and Mi Hee Lim*,†,‡
‡
†Department of Chemistry and Life Sciences Institute, University of Michigan, Ann Arbor, Michigan 48109, United States
∇Department of Fine Chemistry, Seoul National University of Science and Technology, Seoul 139-743, Korea
S
* Supporting Information
ABSTRACT: A diphenylpropynone derivative, DPP2, has been
recently demonstrated to target metal-associated amyloid-β (metal−
Aβ) species implicated in Alzheimer’s disease (AD). DPP2 was shown
to interact with metal−Aβ species and subsequently control Aβ
aggregation (reactivity) in vitro; however, its cytotoxicity has limited
further biological applications. In order to improve reactivity toward
Aβ species and lower cytotoxicity, along with gaining an understanding
of a structure-reactivity-cytotoxicity relationship, we designed,
prepared, and characterized a series of small molecules (C1/C2,
P1/P2, and PA1/PA2) as structurally modified DPP2 analogues. A
similar metal binding site to that of DPP2 was contained in these compounds while their structures were varied to afford different
interactions and reactivities with metal ions, Aβ species, and metal−Aβ species. Distinct reactivities of our chemical family toward
in vitro Aβ aggregation in the absence and presence of metal ions were observed. Among our chemical series, the compound
(C2) with a relatively rigid backbone and a dimethylamino group was observed to noticeably regulate both metal-free and metal-
mediated Aβ aggregation to different extents. Using our compounds, cell viability was significantly improved, compared to that
with DPP2. Lastly, modifications on the DPP framework maintained the structural properties for potential blood-brain barrier
(BBB) permeability. Overall, our studies demonstrated that structural variations adjacent to the metal binding site of DPP2 could
govern different metal binding properties, interactions with Aβ and metal−Aβ species, reactivity toward metal-free and metal-
induced Aβ aggregation, and cytotoxicity of the compounds, establishing a structure-reactivity-cytotoxicity relationship. This
information could help gain insight into structural optimization for developing nontoxic chemical reagents toward targeting
metal−Aβ species and modulating their reactivity in biological systems.
In order to elucidate potential associations of metal−Aβ
species with AD pathogenesis, recent efforts have been made
toward the development of chemical reagents to target metal−
Aβ species, control the interactions between metal ions and Aβ
species, and alter their reactivities (e.g., metal-triggered Aβ
aggregation).3,6,9,10 The structures of these reagents are
rationally selected to achieve both metal chelation and Aβ
interaction (bifunctionality).11−16 Recently, a diphenylpropy-
none derivative (DPP2) was constructed through a rational
structure-based design principle (incorporation approach),
where a metal chelation site was directly installed into an Aβ
interacting framework with minimal structural modifications
(Figure 1).16 DPP2, with a dimethylamino functionality, has
demonstrated the ability to control both metal-free and metal-
induced Aβ aggregation in vitro to different extents; however,
because of its cytotoxicity, it could be only implemented for in
vitro applications.16 Further investigations on the structural
modifications of DPP2 would be valuable to improve its
reactivity toward metal-free and metal-associated Aβ species
INTRODUCTION
■
Alzheimer’s disease (AD) is a global economic and social issue,
currently affecting 5.4 million Americans and an estimated 24
million people worldwide.1 To date, there is no known cure for
AD, and prevailing treatments only offer symptomatic relief for
a short duration of time.2−6 Recently, particular emphasis has
been placed on therapeutics for misfolded amyloidogenic
peptides, amyloid-β (Aβ) peptides, commonly found in AD
brains.2−7 The etiology of AD, however, remains elusive,
because of the complexity emerging from the limited
understanding of pathological features. The role of Aβ itself
and its potential inter-relationships with other multiple factors,
such as metal ion dyshomeostasis and oxidative stress, have
been suggested.2−6,8 Interactions between metal ions (e.g.,
Cu2+, Zn2+) and Aβ have been demonstrated to facilitate
peptide aggregation and enhance oxidative stress via generation
of reactive oxygen species.2,3,6,8,9 The involvement of metal-
associated Aβ (metal−Aβ) species in the onset and progression
of AD, however, has been unclear and contentious, because of
the limited availability of suitable chemical reagents to target
metal−Aβ species and probe their link to neurotoxicity.
Received: April 6, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ic400851w | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 1. A class of DPP derivatives (C1/C2, P1/P2, and PA1/PA2). Chemical structures of chalcone derivative ([125I]-(E)-3-(4-
(dimethylamino)phenyl)-1-(4-iodophenyl)prop-2-en-1-one), DPP2 (3-(4-(dimethylamino)phenyl)-1-(pyridin-2-yl)prop-2-yn-1-one), L2-b
(N1,N1-dimethyl-N4-(pyridin-2-ylmethyl)benzene-1,4-diamine), C1 ((E)-3-phenyl-1-(pyridin-2-yl)prop-2-en-1-one), C2 ((E)-3-(4-
(dimethylamino)phenyl)-1-(pyridin-2-yl)prop-2-en-1-one), P1 (3-phenyl-1-(pyridin-2-yl)propan-1-one), P2 (3-(4-(dimethylamino)phenyl)-1-
(pyridin-2-yl)propan-1-one), PA1 (N-phenylpicolinamide), and PA2 (N-(4-(dimethylamino)phenyl)picolinamide). Atoms for potential metal
binding are highlighted in blue. The green box indicates metal chelation site from DPP2.
128.8, 126.9, 122.9, 120.8. HRMS: Calcd for [M+H]+, 210.0841;
and cytotoxicity, as well as to establish a relationship between
found for [M+H]+, 210.0913.
Preparation of (E)-3-(4-(Dimethylamino)phenyl)-1-(pyridin-
2-yl)prop-2-en-1-one (C2). A flask was charged with a mixture of 2-
acetylpyridine (242 mg, 2.0 mmol) and 4-dimethylaminobenzaldehyde
(300 mg, 2.0 mmol) in EtOH (5 mL). To the stirring mixture was
added a 20% solution of KOH (3 mL) dropwise. After stirring the
chemical structures, reactivity with Aβ and metal−Aβ species,
and toxicity in cells. Thus, we present the design, preparation,
characterization, and in vitro metal-free and metal-associated
Aβ reactivity of a structurally modified DPP2 series (C1/C2,
P1/P2, and PA1/PA2; see Figure 1). The results and
observations presented in this study illustrate that our structural
variations on DPP2 were shown to (i) tune metal binding and
Aβ interaction properties of small molecules, thus altering their
influence on both metal-free and metal-induced Aβ aggregation
in vitro; (ii) reduce their cytotoxicity; and (iii) maintain their
potential BBB permeability.
mixture overnight at room temperature, H2O (5 mL) was added to the
flask before the red precipitates were filtered and washed with H2O (5
mL) and cold EtOH (5 mL). The product was concentrated in vacuo
and purified by recrystallization from CH2Cl2 and hexanes to afford
1
dark orange crystals (352 mg, 1.4 mmol, 70%). H NMR (400 MHz,
CDCl3)/δ (ppm): 3.05 (6H, s), 6.07 (2H, J = 9.2 Hz), 7.46 (1H, m),
7.64 (2 H, J = 9.2 Hz), 7.86 (1H, td, J = 7.6 and 2.0 Hz), 7.94 (1H, d, J
= 16.0 Hz), 8.08 (1H, d, J = 16.0 Hz), 8.19 (1 H, d, J = 8.0 Hz), 8.73
(1 H, m). 13C NMR (100 MHz, CDCl3)/δ (ppm): 189.2, 155.0, 152.1,
148.7, 146.0, 136.9, 130.9, 126.4, 123.0, 122.8, 115.5, 111.7, 40.1.
HRMS: Calcd for [M+H]+, 253.1263; found for [M+H]+, 253.1335.
Preparation of 3-Phenyl-1-(pyridin-2-yl)propan-1-one (P1).
A flame-dried flask was charged with (E)-3-phenyl-1-(pyridin-2-
yl)prop-2-en-1-one (300 mg, 0.71 mmol) and purged for 5 min
prior to addition of dry CH3OH (9 mL). To the stirring solution was
added 10 wt % Pd/C (30 mg, 10 mol %) and Ph2S (2.5 μL, 7.0 μmol).
The reaction was allowed to stir under H2 at room temperature for 24
h. The solution was then filtered through a 25-mm syringe filter,
washed with CH3OH (3 × 10 mL), and dried over anhydrous MgSO4.
The product was concentrated in vacuo and purified by column
chromatography (SiO2, 100% CH2Cl2) to afford an orange oil (108
EXPERIMENTAL SECTION
■
Materials and Procedures. All reagents were purchased from
commercial suppliers and used as received unless otherwise stated.
The compounds (C1/C2 and PA1/PA2) were synthesized following
previously reported methods with modifications.17−20 Aβ1−40
(DAEFRHDSGYEVHHQKLVFFAEDVGSNKGAIIGLMVGGVV)
was purchased from Anaspec (Fremont, CA). Optical spectra for metal
binding and solution speciation studies were recorded on an Agilent
8453 UV−visible (UV−vis) spectrophotometer. Nuclear magnetic
resonance (NMR) spectra for the characterization of small molecules
and Zn2+ binding studies were acquired on a Varian 400 MHz NMR
spectrometer. Isothermal calorimetric titrations of Aβ1−40 with
compounds were carried out with a MicroCal VP-ITC isothermal
titration calorimeter (Northampton, MA). Transmission electron
microscopy (TEM) images were recorded on a Philips CM-100
transmission electron microscope. A SpectraMax M5 microplate
reader (Molecular Devices, Sunnyvale, CA) was used to measure the
absorbance for MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetra-
zolium bromide) and Parallel Artificial Membrane Permeability Assay
adapted for Blood-Brain Barrier (PAMPA-BBB) assays.
1
mg, 0.51 mmol, 72%). H NMR (400 MHz, CDCl3)/δ (ppm): 3.07
(2H, t, J = 7.6 Hz), 3.57 (2H, t, J = 7.6 Hz), 7.18 (1H, q), 7.25−7.26
(4H, br), 7.45−7.42 (1H, m), 7.81 (1H, td, J = 9.2, 1.6 Hz), 8.02 (1H,
dt, J = 8.0, 1.2 Hz), 8.64 (1H, d, J = 4.8 Hz). 13C NMR (100 MHz,
CDCl3)/δ (ppm): 200.8, 153.4, 148.9, 141.7, 136.9, 128.4, 128.3,
127.1, 125.9, 121.5, 39.2, 29.8. HRMS: Calcd for [M+H]+, 212.1070;
found for [M+H]+, 212.1070.
Preparation of 3-(4-(Dimethylamino)phenyl)-1-(pyridin-2-
yl)propan-1-one (P2). A flame-dried flask was charged with (E)-3-
(4-(dimethylamino)phenyl)-1-(pyridin-2-yl)prop-2-en-1-one (300 mg,
0.72 mmol) and purged for 5 min prior to addition of dry CH3OH (9
mL). To the stirring solution was added 10 wt % Pd/C (30 mg, 10
mol %) and Ph2S (2.5 μL, 7.0 μmol). The reaction was allowed to stir
under H2 at room temperature for 24 h. The solution was then filtered
through a 25-mm syringe filter, washed with CH3OH (3 × 10 mL),
and dried over anhydrous MgSO4. The product was concentrated in
vacuo and purified by column chromatography (SiO2, 1:1 ethyl
acetate:hexanes) and washed several times with hexanes to afford the
final product as a light green solid (93 mg, 0.37 mmol, 51%). 1H NMR
(400 MHz, CD2Cl2)/δ (ppm): 2.89 (6H, s), 2.93 (2H, t, J = 8.0 Hz),
Preparation of (E)-3-Phenyl-1-(pyridin-2-yl)prop-2-en-1-one
(C1). A flask was charged with a mixture of 2-acetylpyridine (300 mg,
2.5 mmol) and benzaldehyde (285 mg, 2.7 mmol) in H2O (15 mL).
To the stirring mixture was added a 10% solution of NaOH (1.5 mL)
dropwise. The mixture was stirred overnight at room temperature. The
precipitates formed were filtered, washed with H2O (3 × 10 mL), and
dried over anhydrous MgSO4. The product was concentrated in vacuo
and purified by column chromatography (SiO2, 100% CH2Cl2) to
1
afford a yellow solid (450 mg, 2.2 mmol, 85%). H NMR (400 MHz,
CDCl3)/δ (ppm): 7.42 (3H, m), 7.49 (1H, m), 7.73 (2H, m), 7.88 (1
H, dt, J = 7.6 and 1.6), 7.95 (1H, d, J = 16.0 Hz), 8.20 (1H, d, J = 8.0
Hz), 8.32 (1H, d, J = 16.0 Hz), 8.75 (1 H, m). 13C NMR (100 MHz,
CDCl3)/δ (ppm): 189.4, 154.1, 148.8, 144.8, 137.0, 135.1, 130.5,
B
dx.doi.org/10.1021/ic400851w | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
3.48 (2H, t, J = 8.0 Hz), 6.67 (2H, d, J = 8.0 Hz), 7.11 (2H, d, J = 8.0
Hz), 7.47 (1H, ddd, J = 12.0, 4.0, 1.0 Hz), 7.83 (1H, td, J = 7.6, 1.6
Hz), 7.99 (1H, dt, J = 7.6, 0.8 Hz), 8.65 (1H, dd, J = 4.5, 1.2 Hz). 13C
NMR (100 MHz, CD2Cl2)/δ (ppm): 201.2, 153.5, 149.2, 148.9, 136.8,
129.8, 128.9, 127.1, 121.5, 112.9, 40.7, 39.6, 28.9. HRMS: Calcd for
[M+H]+, 255.1492; found for [M+H]+, 255.1496.
conjugated goat antimouse secondary antibody (1:5,000; Cayman
Chemical, Ann Arbor, MI) in 2% BSA in TBS-T solution for 1 h at
room temperature. The protein bands were visualized using Thermo
Scientific Supersignal West Pico Chemiluminescent Substrate (Rock-
ford, IL). Note that the gel analysis presented herein is qualitative due
to properties of resulting Aβ species (e.g., sizes, conformations).
Transmission Electron Microscopy (TEM). Samples for TEM
were prepared following a previously reported method.12,13,21 Glow-
discharged grids (Formar/Carbon 300-mesh, Electron Microscopy
Sciences, Hatfield, PA) were treated with samples from either
inhibition or disaggregation experiments (5 μL) for 2 min at room
temperature. Excess sample was removed with filter paper and the
grids were washed with ddH2O five times. Each grid was stained with
uranyl acetate (1% w/v ddH2O, 5 μL) for 1 min. Uranyl acetate was
blotted off and grids were dried for 15 min at room temperature.
Images of samples were taken by a Philips Model CM-100
transmission electron microscopy (TEM) system (80 kV, 25 000×
magnification).
Preparation of N-Phenylpicolinamide (PA1). A dried flask was
charged with picolinic acid (123 mg, 1.0 mmol), followed by the
addition of dry CH2Cl2 (10 mL). To the stirring mixture was added
dicyclohexylcarbodiimide (DCC, 230 mg, 2.2 mmol), aniline (0.2 mL,
2.2 mmol), and 4-dimethylaminopyridine (DMAP, 50 mg, 0.45
mmol). After the mixture was stirred overnight at room temperature,
H2O (5 mL) was added to the flask before the white precipitates were
filtered. The organic mixture was washed with H2O (3 × 10 mL) and
dried over anhydrous MgSO4. The product was concentrated in vacuo
and purified by column chromatography (SiO2, 6:1 CH2Cl2:ethyl
1
acetate) to afford a white solid (147 mg, 0.74 mmol, 74%). H NMR
(400 MHz, CDCl3)/δ (ppm): 7.13 (1H, t, J = 7.4 Hz), 7.40 (2H, dd, J
= 7.8, 8.1 Hz), 7.46 (1H, m), 7.78 (2H, d, J = 7.8 Hz), 7.89 (2H, dt, J
= 1.8, 7.8 Hz), 8.29 (1H, d, J = 7.8 Hz), 8.60 (1H, d, J = 4.8 Hz). 13C
NMR (100 MHz, CDCl3)/δ (ppm): 161.9, 149.8, 147.9, 137.7, 129.1,
126.4, 124.3, 122.4, 119.7. HRMS: Calcd for [M+H]+, 199.0793;
found for [M+H]+, 199.0866.
Metal Binding Studies. The interactions of C1/C2, P1/P2, and
1
PA1/PA2 with Cu2+ and Zn2+ were investigated by UV−vis and H
NMR spectroscopy, respectively, based on previously reported
procedures.12,13,21 A solution of C1/C2 (20 μM), P1/P2 (50 μM),
or PA1/PA2 (50 μM) in EtOH was treated with 1−5 equiv of CuCl2
with incubation of 2 min (for C1/C2) or 10 min (for P1/P2 or PA1/
PA2) at room temperature and was monitored by UV−vis. The
interaction of C1/C2, P1/P2, or PA1/PA2 with ZnCl2 was observed
by 1H NMR spectroscopy upon the addition of 1−3 equiv of ZnCl2 to
a solution of C1 (2 mM), C2 (4 mM), P1/P2 (4 mM), or PA1/PA2
(8 mM) in acetonitrile-d3 (CD3CN). In addition, metal selectivity of
C1/C2, P1/P2, and PA1/PA2 was examined by measuring the optical
changes upon addition of 1 equiv of CuCl2 to a solution of ligand
([C1/C2] = 20 μM or [P1/P2] = [PA1/PA2] = 50 μM in EtOH)
containing 1 or 25 equiv of another divalent metal chloride salt
(MgCl2, CaCl2, MnCl2, FeCl2, CoCl2, NiCl2, or ZnCl2). The Fe2+
samples were prepared by purging the solutions with N2.
Quantification of metal selectivity was calculated by comparing and
normalizing the absorption values of metal−ligand complexes at 400
nm (for C1), 650 nm (for C2), 320 nm (for P1), 300 nm (for P2),
350 nm (for PA1), or 480 nm (for PA2) to the absorption at these
wavelengths before and after the addition of CuCl2 (AM/ACu).
Solution Speciation Studies. The pKa values for C1/C2, P1/P2,
and PA1/PA2 were determined by UV−vis variable-pH titrations
following previously reported methods.11,13,16 To obtain the pKa
values, a solution (10 mM NaOH, pH 12, 100 mM NaCl) of C1/
C2 (20 μM), P1/P2 (50 μM), or PA1/PA2 (50 μM) was titrated with
small aliquots of HCl to obtain at least 30 spectra in the range of pH
2−10 (for C1/C2 and P1/P2) or pH 1−12 (for PA1/PA2). In order
to investigate the binding properties of ligands to Cu2+ at various pH
values, a solution containing a compound ([C2] = 20 μM, [P2] = 50
μM, or [PA2] = 50 μM) and CuCl2 in a ratio of 2:1 was incubated for
30 min, 2 h, and 30 min, respectively, and titrated with additions of
HCl in a similar manner. At least 30 spectra were obtained over the
range of pH 2−7 (for C2 and P2) or pH 1−7.5 (for PA2). The acidity
(pKa) and stability (log β) constants were calculated by using the
HypSpec program (Protonic Software, U.K.).25 The speciation
diagrams for C1/C2, P1/P2, PA1/PA2, Cu2+−C2, Cu2+−P2, and
Cu2+−PA2 complexes were modeled by the HySS2009 program
(Protonic Software).26
Preparation of N-(4-(Dimethylamino)phenyl)picolinamide
(PA2). A dried flask was charged with picolinic acid (123 mg, 1.0
mmol), followed by the addition of dry CH2Cl2 (10 mL). To the
stirring mixture was added DCC (230 mg, 2.2 mmol), N1,N1-
dimethylbenzene-1,4-diamine (149 mg, 2.2 mmol), and DMAP (50
mg, 0.45 mmol). After the mixture was stirred overnight at room
temperature, H2O (5 mL) was added to the flask before the white
precipitates were filtered. The organic mixture was washed with water
(3 × 10 mL) and dried over anhydrous MgSO4. The product was
concentrated in vacuo and purified by column chromatography (SiO2,
5:1 CH2Cl2:ethyl acetate) to afford a yellow solid (190 mg, 0.79 mmol,
79%). 1H NMR (400 MHz, CDCl3)/δ (ppm): 2.93 (6H, s), 6.75 (2H,
d, J = 9.0 Hz), 7.43 (1H, m), 7.63 (2H, d, J = 9.0 Hz), 7.86 (1H, dt, J =
1.7, 7.7 Hz), 8.27 (1H, d, J = 7.8 Hz), 8.58 (1H, d, J = 4.8 Hz), 9.82
(1H, s). 13C NMR (100 MHz, DMSO)/δ (ppm): 162.0, 150.7, 148.7,
147.9, 138.5, 128.5, 127.0, 122.5, 121.8, 112.9, 40.9. HRMS: Calcd for
[M+H]+, 242.1205; found for [M+H]+, 242.1292.
Amyloid-β (Aβ) Peptide Experiments. Aβ1−40 peptide was
dissolved with ammonium hydroxide (NH4OH, 1% v/v, aq),
aliquoted, lyophilized, and stored at −80 °C. A stock solution (ca.
200 μM) was prepared by redissolving Aβ with NH4OH (1% w/v, aq,
10 μL) followed by dilution with doubly distilled (dd) H2O, as
reported previously.12,13,21 Buffered solutions (20 μM HEPES (4-(2-
hydroxyethyl)-1-piperazineethanesulfonic acid), pH 6.6 or 7.4, 150 μM
NaCl) were used for both inhibition and disaggregation studies (pH
6.6 for Cu2+ samples;22−24 pH 7.4 for metal-free and Zn2+ samples).
For the inhibition experiment, Aβ (25 μM) was first treated with or
without a metal chloride salt (CuCl2 or ZnCl2, 25 μM) for 2 min
followed by addition of C1/C2, P1/P2, or PA1/PA2 (50 μM, 1% v/v
final DMSO concentration). The resulting samples were incubated at
37 °C for 4, 8, or 24 h with constant agitation. For the disaggregation
experiment, Aβ in the absence or presence of a metal chloride salt
(CuCl2 or ZnCl2) were initially incubated at 37 °C for 24 h with
steady agitation. The compound was added afterward, followed by an
additional 4, 8, or 24 h of incubation at 37 °C with constant agitation.
Gel Electrophoresis with Western Blotting. The Aβ peptide
experiments described above were analyzed by gel electrophoresis,
followed by Western blotting, using an anti-Aβ antibody
(6E10).12,13,21 Each sample (10 μL, [Aβ] = 25 μM) was separated
using a 10−20% gradient Tris-tricine gel (Invitrogen, Grand Island,
NY). The gel was transferred to a nitrocellulose membrane and
blocked overnight with a bovine serum albumin (BSA) solution (3%
w/v, Sigma, St. Louis, MO) in Tris-buffered saline (TBS, Fisher,
Pittsburgh, PA) containing 0.1% Tween-20 (Sigma; TBS-T). The
membrane was treated with the Aβ monoclonal antibody (6E10,
Covance, Princeton, NJ; 1:2000; BSA, 2% w/v, in TBS-T) for 4 h at
room temperature and then probed with a horseradish peroxidase-
Isothermal Titration Calorimetry (ITC). Solutions of ligand (200
μM) and Aβ1−40 (20 μM) in 20 μM HEPES buffer, pH 7.4, 150 μM
NaCl (10% v/v DMSO) were prepared and degassed for 10 min prior
to titration. The ligand solution (10 μL per injection) was injected
over 1 s (25 times, with an interval of 200 s between each injection)
into a solution of Aβ1−40 (1.4 mL) using a motor-driven 250-μL
syringe rotating at 310 rpm at 25 °C. As the control experiment, the
ligand solution was injected into buffer solution without Aβ to
measure the heat of dilution. Heat values of binding were measured by
subtracting the heat of dilution value from the experimental results.
Titration data were analyzed by using evaluation software (MicroCal
Origin, Version 7.0). The binding curves were fitted with a one-site
C
dx.doi.org/10.1021/ic400851w | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
binding or sequential model, affording the overall heat values of
thermodynamic parameters.27
first retained in these frameworks (Figure 1). Metal chelation of
C1/C2 and P1/P2 could occur through nitrogen and oxygen
(N/O) donor atoms from 2-pyridyl ketone, while in PA1/PA2,
N/O (2-pyridyl ketone) or N/N (pyridine and amide) donor
atoms may be involved in metal binding.34 Second, the triple
bond with a carbonyl group (ynone) in DPP2, which may act
as a Michael acceptor and cause cytotoxicity,35−37 was modified
(i.e., reduction or derivation of the triple bond; see Figure 1).
This structural alteration could vary overall structural rigidity.
C1/C2 were produced via reduction of the triple bond to a
double bond; the backbone of C1/C2 was also found in the
framework of an Aβ imaging agent, which is a chalcone
derivative (Figure 1).38 For P1/P2, the double bond was
further reduced to a single bond. PA1/PA2 were fashioned by
combining the 2-pyridyl ketone moiety from DPP2 and the
aniline portion from L2-b via an amide moiety, instead of the
carbon linker shown in DPP2, C1/C2, and P1/P2. Similar to
DPP2, L2-b (Figure 1) was previously designed to serve as a
potential chemical reagent to investigate metal−Aβ species in
vitro and in living cells.13 Lastly, a dimethylamino group, which
has been suggested to be important for Aβ interaction,13,16,39
was incorporated into the frameworks of C2, P2, and PA2. For
preparation, C1/C2 and PA1/PA2 were synthesized based on
previously reported methods with modifications (see Scheme
S1 in the Supporting Information).17−20 P1/P2 were generated
through the selective hydrogenation of C1/C2 in the presence
of 10 wt % Pd/C (see Scheme S1 in the Supporting
Information).
Docking Studies. Flexible ligand docking studies for C1/C2
against the Aβ1−40 monomer from the previously determined aqueous
solution NMR structure (PDB 2LFM)28 were conducted using
AutoDock Vina.29 Ten (10) conformations were selected from 20
conformations within the Protein Databank (PDB) file (1, 3, 5, 8, 10,
12, 13, 16, 17, and 20). The MMFF94 energy minimization in
ChemBio3D Ultra 11.0 was used to optimize the ligand structures for
docking studies. The structural files of C1/C2 and the peptide were
generated by AutoDock Tools and imported into PyRx,30 which were
used to run AutoDock Vina.29 The search space dimensions were set
to contain the entire peptide. The exhaustiveness for the docking runs
was set at 1024. Docked poses of the ligands with Aβ were visualized
using Pymol.
Cytotoxicity (MTT Assay). The human neuroblastoma SK-N-
BE(2)-M17 cell line was purchased from the American Type Culture
Collection (ATCC, Manassas, VA). The cell line was maintained in
media containing 1:1 Minimum Essential Media (MEM, GIBCO,
Grand Island, NY) and Ham’s F12K Kaighn’s Modification Media
(F12K, GIBCO), 10% (v/v) fetal bovine serum (FBS, Sigma), 100 U/
mL penicillin (GIBCO), and 100 mg/mL streptomycin (GIBCO).
The cells were grown and maintained at 37 °C in a humidified
atmosphere with 5% CO2. M17 cells were seeded in a 96-well plate
(15 000 cells in 100 μL per well), according to previously reported
methods.12,13,16,21 These cells were treated with various concentrations
of C1/C2, P1/P2, or PA1/PA2 (2.5−50 μM for C1/C2, 2.5−200 μM
for P1/P2 and PA1/2, final 1% v/v DMSO). After 24 h incubation at
37 °C, 25 μL MTT (5 mg/mL in phosphate buffered saline (PBS), pH
7.4, GIBCO) was added to each well and the plates were incubated for
4 h at 37 °C. Formazan produced by the cells was dissolved in a
solution containing N,N-dimethylformamide (DMF, 50% v/v aq, pH
4.5) and sodium dodecyl sulfate (SDS, 20% w/v) overnight at room
temperature. The absorbance (A600) was subsequently measured on a
microplate reader.
Parallel Artificial Membrane Permeability Assay Adapted
for Blood-Brain Barrier (PAMPA-BBB). PAMPA-BBB experiments
of compounds were carried out using the PAMPA Explorer kit (Pion,
Inc. Billerica, MA) with modification to previously reported
protocols.13,16,31−33 Each stock solution was diluted with pH 7.4
Prisma HT buffer (Pion) to a final concentration of 25 μM (1% v/v
DMSO) and was added to the wells of the donor plate (200 μL,
number of replicates = 12). BBB-1 lipid formulation (5 μL, Pion) was
used to coat the poly(vinylidene fluoride) (PVDF, 0.45 μM) filter
membrane on the acceptor plate. The acceptor plate was placed on top
of the donor plate, forming a sandwich, and brain sink buffer (BSB,
200 μL, Pion) was added to each well of the acceptor plate. The
sandwich was incubated for 4 h at ambient temperature without
stirring. UV−vis spectra of the solutions in the reference, acceptor, and
donor plates were measured on a microplate reader. The PAMPA
Explorer software, v. 3.5 (Pion), was used to calculate the value of
−logPe for each compound. CNS designations were assigned by
comparison to compounds that were identified in previous
reports.31−33
Effects of C1/C2, P1/P2, and PA1/PA2 on Metal-Free
and Metal-Induced Aβ Aggregation In Vitro. In order to
understand the effect of structural variations on compound
reactivity toward metal-free and metal-induced Aβ aggregation,
inhibition (Figure 2) and disaggregation (Figure S1 in the
Supporting Information) studies were conducted. Size
distributions and morphologies of the resulting Aβ species
from the two experiments were monitored by gel electro-
phoresis, followed by Western blot with an anti-Aβ antibody
(6E10) and TEM, respectively.12−16,21
The influence of C1/C2, P1/P2, and PA1/PA2 on
formation of metal-free and metal-treated Aβ aggregates was
first examined (inhibition studies; see Figure 2). Among these
compounds, C2 demonstrated the greatest reactivity with
metal-free and metal-associated Aβ species to different extents,
as confirmed by gel analysis. From metal-free Aβ aggregation
with C2 treatment, a wide molecular weight (MW) distribution
of Aβ species was observed (Figure 2a, left, lane 3), which was
also enhanced with prolonged incubation. Moreover, TEM
images, obtained from the sample incubated with C2 for 24 h
(without metal ions), revealed thinner fibrils with diverse
lengths, compared to those generated under compound-free
conditions (Figure 2b). In the case of metal-triggered Aβ
aggregation, C2 presented differences in the control of Aβ
aggregate formation in the reaction with Cu2+−Aβ over that
with Zn2+−Aβ. Upon treating Cu2+−Aβ species with C2,
various-sized Aβ species were visualized by gel electrophoresis
(Figure 2a, lane 6); while species (MW ≤ 25 kDa) were
exhibited from the sample containing Aβ, Zn2+, and C2 (Figure
2a, lane 9). The modulation of Aβ aggregation by C2 in the
absence and presence of metal ions was also distinguished by
TEM. A mixture of fibrils and amorphous aggregates (mainly,
unstructured Aβ species) was detected in metal-treated
conditions, whereas fibrils were mostly indicated from metal-
free analogues.
RESULTS AND DISCUSSION
■
Design Consideration and Preparation of C1/C2, P1/
P2, and PA1/PA2. A small molecule, DPP2, has been
observed to target Aβ and metal−Aβ species and noticeably
modulate metal-free and metal-induced Aβ aggregation in vitro
to different extents; however, it presented cytotoxicity at low
micromolar concentrations.16 A series of compounds (C1/C2,
P1/P2, and PA1/PA2; see Figure 1) were designed based on
the structural modifications of DPP2 in order to improve its
reactivity with Aβ and metal−Aβ species, mitigate its
cytotoxicity, and establish a structure-reactivity-cytotoxicity
relationship. For the design of the compounds, the structural
portion of DPP2 for metal chelation, 2-pyridyl ketone,16 was
D
dx.doi.org/10.1021/ic400851w | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 2. Inhibition experiment: modulation of Aβ aggregate formation in the absence and presence of metal ions (scheme, top) by C1/C2, P1/P2,
or PA1/PA2: (a) analysis of the size distribution of the resulting Aβ species by gel electrophoresis with Western blot using an anti-Aβ antibody
(6E10); and (b) visualization of the morphologies of the Aβ species from samples that were incubated with C1 or C2 for 24 h by TEM. Conditions:
[Aβ] = 25 μM; [CuCl2 or ZnCl2] = 25 μM; [compound] = 50 μM (where “compound” = C1/C2, P1/P2, or PA1/PA2); pH 6.6 (for Cu2+ samples)
or 7.4 (for metal-free and Zn2+ samples); various inhibition times; 37 °C; constant agitation.
As shown in Figure 2, alterations in structural scaffold
differentiated reactivity toward Aβ aggregation (with and
without metal ions). First, C1, which does not contain a
dimethylamino group on the backbone compared to that of C2
(Figure 1), did not significantly regulate the generation of Aβ
aggregates (Figure 2a, left, lanes 2, 5, and 8). Second, as the
carbon−carbon triple bond (CC) between two structural
moieties (2-pyridyl ketone and dimethylaniline) in DPP216 was
reduced to a double bond (CC) in C2 (Figure 1), the
reactivity was changed toward Aβ aggregation in the absence
and presence of metal ions. Under metal-free conditions, C2
afforded Aβ species with a diverse range of MW starting from
time points of 4−24 h, while DPP2 exhibited less control on
Aβ aggregation as time progressed.16 Cu2+-containing samples
presented Aβ species with a various distribution of MW after
24 h with C2 and DPP2.16 Notably, this reactivity was more
apparent with C2 at early time periods. Aβ aggregates of MW <
25 kDa were indicated over time when treated with Zn2+ and
C2. Aβ species (MW ≤ 25 kDa) were also observed in the case
of DPP2 at a short incubation time (e.g., 4 h), yet species
showing different sizes were yielded after 24 h.16 Third, upon
substitution of the double bond in C2 to a single bond in
P1/P2 (Figure 1), reactivity toward metal-free and metal-
triggered Aβ aggregation disappeared (Figure 2a, middle).
Lastly, for PA1/PA2, composed of 2-pyridyl ketone and an
aniline portion via an amide linkage (Figure 1), some reactivity
was detected toward Cu2+-treated Aβ species after 24 h
of incubation (Figure 2a, right, lanes 5 and 6). In addition,
PA2, which has a dimethylamino functionality, showed
slight modulation on Zn2+-induced Aβ aggregation after 24 h
(Figure 2a, right, lane 9). In contrast to their reactivity toward
metal−Aβ, PA1/PA2 did not display any significant influence
on metal-free Aβ species even with longer incubation time. This
general reactivity of PA1/PA2 was distinct from that of
DPP216 or C2.
Moreover, transformation of preformed metal-free and
metal-induced Aβ aggregates by C1/C2, P1/P2, or PA1/PA2
was also investigated (disaggregation studies; see Figure S1 in
the Supporting Information). Upon incubation of these
compounds, similar size distributions of resulting Aβ aggregates
were confirmed by gel electrophoresis in comparison to those
shown in the inhibition experiment (see Figure 2 and Figure S1
in the Supporting Information). In particular, C2 was able to
noticeably disassemble preformed metal-free and metal-
mediated Aβ aggregates (Figure S1a in the Supporting
Information). Aβ species generated from the disaggregation
experiment employing C2 displayed an array of MW and
morphologies analogous to those in the inhibition studies (see
Figure 2 and Figure S1 in the Supporting Information). The
samples containing metal-treated Aβ species, followed by the
addition of DPP2,16 were visualized to have relatively more
dispersion in MW than those with C2. In the metal-free
condition, however, diverse-sized Aβ species were exhibited
over longer incubation only in the presence of C2. As observed
in inhibition studies, C1 and P1/P2 were not able to influence
the disassembly of preformed Aβ aggregates in both metal-
absent or metal-present cases; the modulation of these peptide
aggregates was indicated with PA1/PA2 only after 24 h. Thus,
structural variations, such as the introduction of flexibility
into the framework and/or incorporation of a dimethylamino
E
dx.doi.org/10.1021/ic400851w | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
moiety, afforded altered reactivity with metal-free and metal-
treated Aβ species in both inhibition and disaggregation studies.
Both experiments suggest that C2, with a double bond
replacing the triple bond in DPP2,16 displayed the greatest
control in regulating Aβ aggregate generation as well as
disassembling preformed Aβ aggregates (both in the absence
and presence of metal ions) to varying degrees in vitro.
Furthermore, the backbones of P and PA, designed by further
modification of the double bond in C2 to a single bond and an
amide bond, respectively, presented no or slight modulation on
the formation or disaggregation of metal-free or metal-induced
Aβ aggregates. Taken together, results and observations from
both inhibition and disaggregation experiments imply that
structural features, such as scaffold rigidity and substituents,
may be essential for directing in vitro reactivity of small
molecules toward metal-free and metal-associated Aβ aggrega-
tion.
optical shifts from ca. 431 and 332 nm to 564 and 408 nm were
detected after CuCl2 treatment of C2 and PA2, respectively. In
addition to Cu2+ coordination, Zn2+ binding of C1/C2, P1/P2,
and PA1/PA2 were studied by 1H NMR (see Figure S2 in the
Supporting Information). NMR was employed for Zn2+ binding
studies due to lack of significant optical changes of these ligands
with ZnCl2. The introduction of 1−3 equiv of Zn2+ to a
CD3CN solution of C1/C2, P1/P2, or PA1/PA2 resulted in
downfield chemical shifts of the pyridyl protons, demonstrating
the involvement of the N donor atom from the pyridine ring
(Figure 1) for Zn2+ binding. Moreover, the α and β protons of
C1/C2 and P1/P2 displayed chemical shifts with Zn2+
treatment, which indicated that the O donor atom of the
carbonyl group may also be associated in Zn2+ binding. In
PA1/PA2, the amide proton shifted upfield and sharpened in
the presence of Zn2+, suggesting that the N donor atom of the
amide moiety may be involved in metal coordination.40,41
Overall, C1/C2, P1/P2, and PA1/PA2, generated via structural
modification from DPP2, were able to bind to metal ions, such
as Cu2+ and Zn2+, possibly through the N and/or O donor
atoms, similar to DPP2.16
Metal Binding Properties of C1/C2, P1/P2, and PA1/
PA2. To investigate the effect of structural variation on metal
binding properties of C1/C2, P1/P2, and PA1/PA2
(containing a similar putative metal chelation site as DPP2,
2-pyridyl ketone16), the ability of C1/C2, P1/P2, or PA1/PA2
To understand solution speciation of our chemical series in
the absence and presence of Cu2+, UV−vis variable-pH titration
experiments were conducted.11,13,16 Spectrophotometric titra-
tions of compounds C1/C2, P1/P2, and PA1/PA2 were first
performed to estimate the acidity constant (pKa) values (see
1
to bind Cu2+ and Zn2+ was first examined by UV−vis and H
NMR spectroscopy, respectively (see Figure 3 and Figure S2 in
the Supporting Information). As shown in Figure 3, changes in
UV−vis spectra (i.e., new optical bands and/or change in
absorbance intensity) were exhibited upon treatment of CuCl2
(1−5 equiv) with C1/C2, P1/P2, or PA1/PA2 in EtOH,
indicating Cu2+ interaction of these compounds. Noticeable
Figure 4; C1, pKa = 2.57(4); C2, pKa1 = 3.23(1), pKa2
=
4.00(1); P1, pKa = 2.59(5); P2, pKa1 = 2.44(2), pKa2 = 5.40(4);
PA1, pKa = 1.69(9); PA2, pKa1 = 1.38(5), pKa2 = 4.91(8)). The
diagrams depicted based on acidity constants indicated the
presence of two forms of ligand (neutral and monoprotonated
species; L and LH; L = C1, P1, and PA1; pKa: protonation of
the pyridyl N), as well as three forms of ligand (neutral,
monoprotonated, and diprotonated species; L, LH, and LH2; L
= C2, P2, and PA2; pKa1: protonation of the pyridyl N, pKa2:
protonation of the dimethylamino N) (Figure 4). The pKa2
values of C2, P2, and PA2 are lower than that of DPP2 (pKa2 =
7.01(6)),16 which implies that neutral species would be present
in a greater magnitude at a physiologically relevant pH (i.e.,
7.4). Overall, all six compounds are expected to mainly exist in
the neutral form at pH 7.4, suggesting their potential
penetration through the blood-brain barrier (BBB; vide
infra).13,16,31−33
Through solution speciation experiments of C2/P2/PA2 in
the presence of CuCl2, their Cu2+ binding stochiometries,
affinities, and the distribution of Cu2+−ligand species were
investigated (see Figure 5).11,13,16 Based on the measured pKa
values of ligands above, the stability constants (log β) of Cu2+−
C2/P2/PA2 complexes were first determined (see Figure 5;
Cu + LH ⇌ Cu(LH) (log β1); Cu + L ⇌ CuL (log β2); Cu +
2L ⇌ CuL2 (log β3); for C2, log β1 = 8.73(0), log β2 = 5.62(8);
for P2, log β1 = 6.50(1), log β2 = 3.76(9); for PA2, log β1
=7.76(9), log β2 = 4.17(7), log β3 = 3.86(5)). Based on the log
β values, greater stability of Cu2+−ligand complexes was shown
in the order of C2, PA2, and P2. Second, solution speciation
diagrams were also modeled using these stability constants,
indicating that (i) complexation occurred between Cu2+ and
ligand in mainly 1:1 ratio (Cu:L; L = C2, P2, or PA2), (ii) free
Cu2+, CuL and Cu(LH) complexes were present in the solution
containing Cu2+ and L (L = C2 or PA2), and (iii) free Cu2+ and
CuL complexes were shown in the sample of P2. At pH 7,42 the
complexes of CuL were present in solution at ca. 80%, 20%,
and 40% for C2, P2, and PA2, respectively. From the diagrams,
Figure 3. Cu2+ binding studies of C1/C2, P1/P2, and PA1/PA2.
UV−vis spectra of (a) C1 and C2, (b) P1 and P2, and (c) PA1 and
PA2 with CuCl2 (1−5 equiv) in EtOH. Conditions: [C1 or C2] = 20
μM; [P1/P2 or PA1/PA2] = 50 μM; room temperature; incubation
for 2 or 10 min.
F
dx.doi.org/10.1021/ic400851w | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 4. Solution speciation studies of C1/C2, P1/P2, and PA1/PA2. UV−vis variable-pH titration spectra (left) and solution speciation diagrams
(right) of (a) C1 and C2 (pH 2−10), (b) P1 and P2 (pH 2−10), (c) PA1 and PA2 (pH 1−10) (FL = fraction of species at the given pH). Acidity
constants (pKa) of L (L = C1/2, P1/P2, or PA1/PA2) are summarized in the table. Conditions: [L] = 20 μM (L = C1 or C2) or 50 μM (L = P1,
P2, PA1, or PA2); I = 0.10 M NaCl; room temperature. Charges are omitted for clarity. The superscript “a” in the table portion of the figure
indicates that the error in the last digit is shown in the parentheses.
the values of pCu (pCu = −log[Cuunchelated]) of these
compounds were also calculated to be 5.79, 4.70, and 4.81 at
pH 6.622−24 for C2, P2, and PA2, respectively. Thus, the
approximate dissociation constants (Kd ≈ [Cuunchelated]) for C2,
P2, and PA2 with Cu2+ were predicted to be in the micromolar
range, making them slightly weaker metal chelators, compared
to DPP2 (pCu = 6.6, ca. low micromolar).16 Different Cu2+
binding affinities of our ligands could result from the influence
of electron density on the donor atoms for metal binding
through structural variations (e.g., structural moieties between
the 2-pyridyl ketone and aniline portions). Higher binding
ability of DPP2 and C2 for Cu2+, in comparison with those of
P2 and PA2, as well as Aβ interaction properties, may direct
their ability to modulate Cu2+-induced Aβ aggregation.
Furthermore, potential selectivity of C2, P2, and PA2 for
Cu2+ over other biologically relevant divalent metal ions (Mg2+,
Ca2+, Mn2+, Fe2+, Co2+, Ni2+, and Zn2+) was studied by
competition reactions which were monitored by UV−vis. As
depicted in Figure S3 in the Supporting Information, at an
equimolar concentration of another metal ion and Cu2+, the
spectral changes after addition of Cu2+ indicated possible
binding of compounds to Cu2+ in the presence of other divalent
metal ions, similar to DPP2.16 Moreover, these results may
suggest a slightly higher selectivity for Cu2+ by C2, P2, and PA2
in the solution containing CoCl2 or NiCl2, compared to
DPP2.16 At a supramolar ratio of another divalent metal ion to
Cu2+ (25:1), P2 seemed to be more selective for Cu2+ than C2
and PA2.
Interaction of C1/C2 with Aβ Species Monitored by
Isothermal Titration Calorimetry (ITC). In order to shed
some light on possible interactions of C1/C2, P2, PA2, and
DPP2 with Aβ1−40, isothermal titration calorimetry (ITC)
experiments were performed, following previously reported
methods.27 Measured heat changes during the titration
indicated potential interaction of our compounds with Aβ.43
The titration data for C1 and C2 were fitted using a one-site
binding model, whereas the data for DPP2 were fitted with a
sequential binding model (see Figure S4 in the Supporting
Information). Based on the binding models, the best-fit values
(binding stoichiometries (N), binding constants (KA), enthalpy
changes (ΔH), and Gibbs free energy (ΔG)) of these
molecules were determined (see Table 1). The binding
G
dx.doi.org/10.1021/ic400851w | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
affinities of all three ligands with Aβ were estimated to be in the
same range (ca. 104−105 M−1). The higher binding stoichio-
metry of C2 to Aβ, compared to that of C1, however, implied
that C2 may be able to interact with Aβ to a greater extent (see
Table 1). This may explain the better reactivity of C2 toward
Aβ species over C1 observed from Aβ aggregation studies (vide
supra; see Figure 2 and Figure S1 in the Supporting
Information). DPP2 also presented similar binding affinity
toward Aβ as C2. Overall, C2 and DPP2 were shown to interact
with Aβ in an energetically favorable manner (see Table 1).
Docking Studies for Possible Conformations of C1/C2
with Aβ. To visualize and further explore possible interactions
between Aβ and C1 or C2, docking studies were carried out
using AutoDock Vina,29 utilizing the previously determined
NMR structure (PDB 2LFM)28 of monomeric Aβ1−40 (see
Figure S5 in the Supporting Information). Aβ and C1/C2 were
observed to interact with each other through nonspecific and/
or direct intermolecular interactions (e.g., hydrogen bonding,
π−π stacking), with most conformations that exhibited C1/C2
docking between the α-helix and the N-terminal random coil,
similar to DPP2.16 These interactions had calculated binding
energies ranging from −5.8 to −4.9 kcal/mol and −5.9 to −4.8
kcal/mol for C1 and C2, respectively (see Figure S5 in the
Supporting Information). As shown in all conformations, C1 and
C2 did not dock at the same sites on Aβ and were situated offset
from each other when docked at similar sites (see Figure S5 in
the Supporting Information). The observations from preliminary
docking studies suggest that C1 and C2 could interact differently
with Aβ, possibly related to the presence of a dimethylamino
group in C2. This may further support the different reactivity
toward Aβ aggregation of C1 and C2 in vitro (vide supra; see
Figure 2 and Figure S1 in the Supporting Information).
Cytotoxicity of C1/C2, P1/P2, and PA1/PA2 in Living
Cells. Cytotoxicity studies of C1/C2, P1/P2, and PA1/PA2
have illustrated a potential relationship between molecular
scaffolds and cytotoxicity. Human neuroblastoma SK-N-BE(2)-
M17 cells were treated with various concentrations of
compounds, and cell viability was measured using the MTT
assay (see Figure 6).13,16,21 Overall, through reduction of the
conjugated triple bond to a double bond in C1/C2, cell
viability (ca. 60% viability at 50 μM) was improved compared
to that of DPP2 (ca. 20% viability at 50 μM).16 Furthermore,
complete removal of the α,β-unsaturated moiety in P1/P2
and PA1/PA2 significantly mitigated the cytotoxicity of com-
pounds (ca. 80% cell viability up to 200 μM concentration).
Through structural modifications, we were able to improve the
cytotoxicity of DPP2.44
Figure 5. Solution speciation studies of the Cu2+−C2, Cu2+−P2, and
Cu2+−PA2 complexes. UV−vis variable-pH spectra (left) and solution
speciation diagrams (right) of (a) Cu2+−C2 (pH 2−7), (b) Cu2+−P2
(pH 2−7), and (c) Cu2+−PA2 (pH 1−7) (FCu = fraction of free Cu
and Cu-L complexes at the given pH). Conditions: [Cu2+]/[L] = 1:2,
[L] = 20 μM (L = C2) or 50 μM (L = P2 or PA2); incubation of
ligand with Cu2+ for 30 min (L = C2 and PA2) or 2 h (L = P2) prior
to pH titration; room temperature. Stability constants (logβ) of Cu−L
complexes are summarized in the table. Charges are omitted for clarity.
The superscript “a” in the table portion of the figure indicates that the
error in the last digit is shown in the parentheses.
Figure 6. Cytotoxicity studies of C1/C2, P1/P2, and PA1/PA2 in human neuroblastoma SK-N-BE(2)-M17 cells. Cells were treated with various
concentrations of (a) C1/C2 (2.5−50 μM), (b) P1/P2 (2.5−200 μM), and (c) PA1/PA2 (2.5−200 μM) for 24 h of incubation. Cell viability was
determined by the MTT assay. Values of cell viability (%) were calculated relative to that of cells incubated only with 1% v/v DMSO. Error bars
represent the standard deviation from at least three independent experiments.
H
dx.doi.org/10.1021/ic400851w | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Table 1. Thermodynamic Parameters of Interactions between C1/C2 or DPP2 with Aβ1‑40 at 25 °C
a
parameter
C1
C2
DPP2
binding stoichiometry, N
0.88 0.27
2.12 0.06
−
4.06 1.00 × 105
7.27 1.26 × 104
3.97 0.80 × 104
binding constant, KA (M−1
)
2.38 0.56 × 105
1.61 0.26 × 105
12.87 0.21 × 105
33.70 2.52 × 104
−15.18 3.70 × 104
enthalpy change, ΔH (kJ/mol)
15.48 9.16
10.93 0.58
−32.01 0.61
−27.75 0.43
−26.25 0.50
Gibbs free-energy change, ΔG (kJ/mol)
−0.69 0.58
−29.72 0.40
a
Thermodynamic values for DPP2 were calculated using a sequential binding mode.
Table 2. Values (MW, clogP, HBA, HBD, PSA, logBB, and −logPe) of C1/C2, P1/P2, and PA1/PA2
a
calculation
C1
209
C2
252
P1
211
P2
254
PA1
198
PA2
241
DPP2
250
Lipinski’s rules and others
molecular weight, MW
≤450
≤5.0
calculated logarithm of the octanol−
water partition coefficient, clogP
2.92
3.09
2.98
3.14
1.26
1.43
2.63
hydrogen-bond acceptor atoms, HBA
hydrogen-bond donor atoms, HBD
polar surface area, PSA
2
3
2
3
3
4
3
≤10
0
0
0
0
1
1
0
≤5
30.0
0.131
33.2
0.108
30.0
0.139
33.2
0.116
42.0
−0.300
45.2
−0.323
33.2
0.0390
≤90 Å2
b
logBB
less than −1.0 (poorly
distributed in the brain)
c
−logPe
4.3 0.1
CNS+
4.2 0.1
4.4 0.1
CNS+
4.3 0.1
CNS+
4.2 0.1
4.2 0.1
4.2 0.1 −logPe < 5.4 (CNS+)
−logPe > 5.7 (CNS−)
d
CNS prediction
CNS+
CNS+
CNS+
CNS+
a
b
c
Values for DPP2 have been previously reported.16 logBB = −0.0148 × PSA + 0.152 × clogP 0.130. −logPe values were determined using the
d
Parallel Artificial Membrane Permeability Assay adapted for BBB (PAMPA-BBB). Prediction of the ability of the compound to penetrate the central
nervous system (CNS). Compounds categorized as CNS+ have the ability to permeate through the BBB and target the CNS. In the case of
compounds assigned as CNS−, they have poor permeability through the BBB and, therefore, their bioavailability into the CNS is considered to be
minimal.
Predicted Blood-Brain Barrier (BBB) Permeability of
C1/C2, P1/P2, and PA1/PA2. To probe whether the scaffold
modifications in the six compounds could retain structural
parameters required for BBB permeability predicted with
DPP2,16 the values of Lipinski’s rules and logBB were first cal-
culated.31,32,45,46 Based on the calculated values (see Table 2),
all six compounds indicated drug-likeness and possible brain
uptake. To confirm the potential BBB penetration prediction,
an in vitro PAMPA-BBB assay was performed, following a pre-
viously reported method.16,31−33 Based on the empirical classifi-
cation of BBB-permeable molecules (e.g., verapamil),16,31−33
the measured permeability values (−logPe), along with neutral
speciation state at a physiologically relevant pH (vide supra; see
Figure 4), suggest that C1/C2, P1/P2, and PA1/PA2 could
potentially permeate through the BBB and be available in the
central nervous system (CNS+). Taken together, our structural
alterations to DPP2 did not seem to change the structural
properties necessary for BBB permeability.
distinct. Particularly, C2, having a double bond between two
structural moieties (2-pyridyl ketone and dimethylaniline) and a
dimethylamino group, displayed noticeable reactivity with both
metal-free and metal-treated Aβ species. The difference in reactivity
was associated with metal binding and Aβ interaction properties of
our compounds, which were confirmed by solution speciation and
ITC studies. In addition, an increase in cell viability was indicated
with C1/C2, P1/P2, and PA1/PA2, relative to that with DPP2,
possibly due to the alteration of the α,β-ynone moiety. Taken
together, through structural modifications of DPP2, we
demonstrate a relationship between structures, reactivity with Aβ
and metal−Aβ, and cytotoxicity. The reactivity of our compounds
with Aβ and metal−Aβ species, however, is shown to be limited in
vitro;44 therefore, further structural optimization for interacting
with metal ions and/or Aβ species is desirable to accomplish their
use in biological systems. Especially, the metal binding affinity of C,
P, and PA could be improved through substitution of the carbonyl
group in their backbones with a stronger metal binding group, such
as an amino or hydroxyl moiety, as well as variation of denticity.
CONCLUSION
■
DPP2 derivatives, C1/C2, P1/P2, and PA1/PA2, were
synthesized in order to improve reactivity of DPP2 toward
Aβ and metal−Aβ species and lower its cytotoxicity, as well as
establish a structure-reactivity-cytotoxicity relationship. The
ability of these structurally modified compounds to modulate in
vitro metal-free and metal-induced Aβ aggregation was significantly
ASSOCIATED CONTENT
■
S
* Supporting Information
Scheme S1 and Figures S1−S5 are available in the Supporting
Information. This material is available free of charge via the
Internet at http://pubs.acs.org.
I
dx.doi.org/10.1021/ic400851w | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
(20) Sasaki, K.; Crich, D. Org. Lett. 2011, 13, 2256−2259.
AUTHOR INFORMATION
Corresponding Author
*E-mail address: mhlim@umich.edu.
Present Addresses
■
(21) Hyung, S.-J.; DeToma, A. S.; Brender, J. R.; Lee, S.;
Vivekanandan, S.; Kochi, A.; Choi, J.-S.; Ramamoorthy, A.; Ruotolo,
B. T.; Lim, M. H. Proc. Natl. Acad. Sci. U.S.A. 2013, 110, 3743−3748.
(22) Physiological acidosis has been observed in the AD brain (i.e.,
pH 6.6). To correlate diseased conditions, experiments were
performed at a relevant pH. In addition, Aβ aggregation in the
presence of Cu2+ was also suggested to be facilitated in a slightly acidic
environment (see refs 23 and 24).
⊥Department of Molecular and Cellular Biology, University of
Guelph, Guelph, Ontario N1G 2W1, Canada.
○Department of Chemistry, University of Calgary, Calgary, AB
T2N 1N4, Canada.
(23) Yates, C. M.; Butterworth, J.; Tennant, M. C.; Gordon, A. J.
Neurochem. 1990, 55, 1624−1630.
(24) Atwood, C. S.; Moir, R. D.; Huang, X.; Scarpa, R. C.; Bacarra, N.
M. E.; Romano, D. M.; Hartshorn, M. A.; Tanzi, R. E.; Bush, A. I. J.
Biol. Chem. 1998, 273, 12817−12826.
(25) Gans, P.; Sabatini, A.; Vacca, A. Ann. Chim. 1999, 89, 45−49.
(26) Alderighi, L.; Gans, P.; Ienco, A.; Peters, D.; Sabatini, A.; Vacca,
A. Coord. Chem. Rev. 1999, 184, 311−318.
(27) Wang, S.-H.; Liu, F.-F.; Dong, X.-Y.; Sun, Y. J. Phys. Chem. B
2010, 114, 11576−11583.
Author Contributions
§These authors contributed equally.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This study was supported by the Ruth K. Broad Biomedical
Foundation, American Heart Association, and National Science
Foundation (No. CHE-1253155) (to M.H.L.). D.L. thanks the
Basic Science Program, through the National Research
Foundation of Korea (NRF), funded by the Ministry of
Education, Science and Technology (No. 2009-0087836) for
the support. A.K. is grateful for the Research Excellence
Fellowship from the Department of Chemistry at the
University of Michigan. We thank Alaina DeToma and Whitney
Smith for assistance of the PAMPA-BBB assay and the
synthesis of C1 and C2, respectively.
(28) Vivekanandan, S.; Brender, J. R.; Lee, S. Y.; Ramamoorthy, A.
Biochem. Biophys. Res. Commun. 2011, 411, 312−316.
(29) Trott, O.; Olson, A. J. J. Comput. Chem. 2010, 31, 455−461.
(30) Wolf, L. K. Chem. Eng. News 2009, 87, 31.
(31) Di, L.; Kerns, E. H.; Fan, K.; McConnell, O. J.; Carter, G. T.
Eur. J. Med. Chem. 2003, 38, 223−232.
(32) Avdeef, A.; Bendels, S.; Di, L.; Faller, B.; Kansy, M.; Sugano, K.;
Yamauchi, Y. J. Pharm. Sci. 2007, 96, 2893−2909.
(33) BBB Protocol and Test Compounds; Pion, Inc.: Woburn, MA,
2009.
(34) Gomes, L.; Low, J. N.; Valente, M. A. D. C.; Freire, C.; Castro,
B. Acta Cryst. 2007, C63, m293−m296.
REFERENCES
■
(35) Schultz, T. W.; Yarbrough, J. W. SAR QSAR Environ. Res. 2004,
15, 139−146.
(1) Alzheimer’s Association. Alzheimer’s Dementia 2012, 8, 131−168.
(2) Kepp, K. P. Chem. Rev. 2012, 112, 5193−5239.
(3) Scott, L. E.; Orvig, C. Chem. Rev. 2009, 109, 4885−4910.
(4) Jakob-Roetne, R.; Jacobsen, H. Angew. Chem. Int. Ed. 2009, 48,
3030−3059.
(36) Koleva, Y. K.; Madden, J. C.; Cronin, M. T. D. Chem. Res.
Toxicol. 2008, 21, 2300−2312.
(37) Sakagami, H.; Kawase, M.; Wakabayashi, H.; Kurihara, T.
Autophagy 2007, 3, 493−495.
(5) Rauk, A. Chem. Soc. Rev. 2009, 38, 2698−2715.
(6) DeToma, A. S.; Salamekh, S.; Ramamoorthy, A.; Lim, M. H.
Chem. Soc. Rev. 2012, 41, 608−621.
(38) Cui, M.; Ono, M.; Kimura, H.; Liu, B. L.; Saji, H. Bioorg. Med.
Chem. Lett. 2011, 21, 980−982.
(39) Leuma Yona, R.; Mazer
2008, 3, 63−66.
̀
es, S.; Faller, P.; Gras, E. ChemMedChem
(7) Hardy, J. A.; Higgins, G. A. Science 1992, 256, 184−185.
(8) Savelieff, M. G.; Lee, S.; Liu, Y.; Lim, M. H. ACS Chem. Biol.
2013, 8, 856−865.
(40) Xu, Z.; Baek, K.-H.; Kim, H. N.; Cui, J.; Qian, X.; Spring, D. R.;
Shin, I.; Yoon, J. J. Am. Chem. Soc. 2010, 132, 601−610.
(41) Braymer, J. J.; Merrill, N. M.; Lim, M. H. Inorg. Chim. Acta
2012, 380, 261−268.
(9) Pithadia, A. S.; Lim, M. H. Curr. Opin. Chem. Biol. 2012, 16, 67−
73.
(10) Rodríguez-Rodríguez, C.; Telpoukhovskaia, M.; Orvig, C. Coord.
Chem. Rev. 2012, 256, 2308−2332.
(42) Because of limited optical features at high pH (>8), solution
speciation studies of Cu2+−ligand complexes were conducted up to pH
7.
(43) The titration data for P2 and PA2 could not be fitted with any
binding model.
́
(11) Rodríguez-Rodríguez, C.; Sanchez de Groot, N.; Rimola, A.;
́
́
Alvarez-Larena, A.; Lloveras, V.; Vidal-Gancedo, J.; Ventura, S.;
Vendrell, J.; Sodupe, M.; Gonzalez-Duarte, P. J. Am. Chem. Soc. 2009,
131, 1436−1451.
́
(44) Mitigation of Aβ or metal−Aβ-induced toxicity by our
compounds was further investigated; however, no noticeable influence
on cell viability was observed (data not shown).
(45) Lipinski, C. A.; Lombardo, F.; Dominy, B. W.; Feeney, P. J. Adv.
Drug Delivery Rev. 2001, 46, 3−26.
(12) Hindo, S. S.; Mancino, A. M.; Braymer, J. J.; Liu, Y.;
Vivekanandan, S.; Ramamoorthy, A.; Lim, M. H. J. Am. Chem. Soc.
2009, 131, 16663−16665.
(13) Choi, J.-S.; Braymer, J. J.; Nanga, R. P. R.; Ramamoorthy, A.;
Lim, M. H. Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 21990−21995.
(14) Braymer, J. J.; Choi, J.-S.; DeToma, A. S.; Wang, C.; Nam, K.;
Kampf, J. W.; Ramamoorthy, A.; Lim, M. H. Inorg. Chem. 2011, 50,
10724−10734.
(46) Clark, D. E.; Pickett, S. D. Drug Discovery Today 2000, 5, 49−58.
(15) Sharma, A. K.; Pavlova, S. T.; Kim, J.; Finkelstein, D.; Hawco, N.
J.; Rath, N. P.; Kim, J.; Mirica, L. M. J. Am. Chem. Soc. 2012, 134,
6625−6636.
(16) Pithadia, A. S.; Kochi, A.; Soper, M. T.; Beck, M. W.; Liu, Y.;
Lee, S.; DeToma, A. S.; Ruotolo, B. T.; Lim, M. H. Inorg. Chem. 2012,
51, 12959−12967.
(17) Zakharychev, V. V.; Kuzenkov, A. V. Chem. Heterocycl. Compd.
2007, 43, 989−995.
(18) Singh, P. K.; Singh, V. K. Org. Lett. 2008, 10, 4121−4124.
(19) Carpino, L. A.; El-Faham, A. J. Org. Chem. 1995, 60, 3561−
3564.
J
dx.doi.org/10.1021/ic400851w | Inorg. Chem. XXXX, XXX, XXX−XXX