Structure-Activity Relationship Studies of Isomeric 2,4-Diaminoquinazolines on β-Amyloid Aggregation Kinetics

Article abstract of DOI:10.1021/acsmedchemlett.6b00039

A library of isomeric 2,4-diaminoquinazoline (DAQ) derivatives were synthesized and evaluated for antiaggregation potential toward Aβ40/42. Structure-activity relationship data identified compound 3k (N⁴-(4-bromobenzyl)quinazoline-2,4-diamine) with a 4-bromobenzyl substituent as the most potent inhibitor (Aβ40 IC₅₀ = 80 nM) and was almost 18-fold more potent compared to the reference agent curcumin (Aβ40 IC₅₀ = 1.5 μM). The corresponding N²-isomer 4k (N²-(4-bromobenzyl)quinazoline-2,4-diamine) was also able to prevent Aβ aggregation (Aβ40 IC₅₀ = 1.7 μM). However, compound 4k exhibited superior inhibition of Aβ42 aggregation (Aβ42 IC₅₀ = 1.7 μM) compared to compound 3k (Aβ42 IC₅₀ = 14.8 μM) and was ～1.8-fold more potent compared to curcumin (Aβ42 IC₅₀ = 3.1 μM). These results were supported by Aβ aggregation kinetics investigations and transmission electron microscopy studies, which demonstrate the suitability of DAQ ring system to develop antiamyloid agents as pharmacological tools to study Aβ aggregation.

Full text of DOI:10.1021/acsmedchemlett.6b00039

Letter
pubs.acs.org/acsmedchemlett
Structure−Activity Relationship Studies of Isomeric 2,4-
Diaminoquinazolines on β‑Amyloid Aggregation Kinetics
Tarek Mohamed,†,‡ Arash Shakeri,† Gary Tin,† and Praveen P. N. Rao*,†
†School of Pharmacy, Health Sciences Campus, and ‡Department of Chemistry, University of Waterloo, 200 University Avenue West,
Waterloo, Ontario N2L 3G1, Canada
S
* Supporting Information
ABSTRACT: A library of isomeric 2,4-diaminoquinazoline
(DAQ) derivatives were synthesized and evaluated for
antiaggregation potential toward Aβ40/42. Structure−activity
relationship data identified compound 3k (N4-(4-bromobenzyl)-
quinazoline-2,4-diamine) with a 4-bromobenzyl substituent as
the most potent inhibitor (Aβ40 IC50 = 80 nM) and was almost
18-fold more potent compared to the reference agent curcumin
(Aβ40 IC50 = 1.5 μM). The corresponding N2-isomer 4k (N2-
(4-bromobenzyl)quinazoline-2,4-diamine) was also able to
prevent Aβ aggregation (Aβ40 IC50 = 1.7 μM). However,
compound 4k exhibited superior inhibition of Aβ42 aggregation
(Aβ42 IC50 = 1.7 μM) compared to compound 3k (Aβ42 IC50 =
14.8 μM) and was ∼1.8-fold more potent compared to
curcumin (Aβ42 IC50 = 3.1 μM). These results were supported
by Aβ aggregation kinetics investigations and transmission electron microscopy studies, which demonstrate the suitability of
DAQ ring system to develop antiamyloid agents as pharmacological tools to study Aβ aggregation.
KEYWORDS: Quinazolines, amyloid, Aβ aggregation, Alzheimer’s disease
ne of the most influential hallmarks of Alzheimer’s
Opathology is the collapse of cellular amyloid manage-
ment, leading to the progressive accumulation of neurotoxic
Aβ-deposits.1−4 While the mechanisms intertwining the
amyloid pathway are vast and complex, a number of strategies
to combat its neuronal insults have been proposed, researched,
and evaluated in various preclinical and clinical settings.5−7 An
essential tool utilized by researchers to further understand the
aggregation mechanisms of amyloidogenic peptides, such as Aβ,
is to develop and evaluate aggregation modulators and
inhibitors.8,9 Not only is this beneficial in understanding the
kinetics of amyloid aggregation but it may reveal potential
therapeutic candidates.
When it comes to developing drug candidates, nature plays
an important role in highlighting ideas for potential core
templates and scaffolds. Honing in on the case of Aβ with
Figure 1. Structures of naturally occurring and synthetic amyloid
aggregation inhibitors and the DAQ scaffold.
respect to aggregation modulators and inhibitors, compounds
the backbone hydrogen bonding interactions in the beta-sheet
assembly thereby providing a framework to design small
molecules as pharmacological tools to study Aβ aggregation and
inhibition.16,17 As part of our research program aimed at
discovering and developing novel small-molecules as potential
pharmacological tools to study Alzheimer’s disease (AD) and
such as curcumin, a component of the spice turmeric, and
resveratrol, a phytoalexin found in grapes and berries (Figure
1), are considered as model compounds in this context.10−13
The synthetic compound, orange G is a commonly used stain/
dye and, like curcumin and resveratrol, is used as a
pharmacological tool in drug discovery due to its excellent
activity against amyloid aggregation.14,15
From a chemical standpoint, these small molecules share
structural features including aromaticity, conjugation, and
planarity resulting in their ability to intercalate and disrupt
Received: January 26, 2016
Accepted: March 1, 2016
© XXXX American Chemical Society
A
DOI: 10.1021/acsmedchemlett.6b00039
ACS Med. Chem. Lett. XXXX, XXX, XXX−XXX

ACS Medicinal Chemistry Letters
Letter
design novel anti-AD agents, we embarked on the development
of isomeric 2,4-diamino-quinazolines (DAQ) library as a novel
class of compounds that exhibit anti-Aβ aggregation properties.
The planar, bicyclic quinazoline ring template can be
considered as a bioisostere of the naphthalene ring present in
orange G (Figure 1). A library of 34 isomeric DAQs were
synthesized and their anti-Aβ aggregation activity was evaluated
by monitoring the Aβ40 and Aβ42 aggregation kinetics using
ThT-fluorescence and by transmission electron microscopy
(TEM) measurements. Computational experiments were used
to propose their binding interactions with Aβ-aggregates. These
studies show that the anti-Aβ activity was sensitive to isomeric
placement of substituents either at the 2 or 4-position of the
quinazoline amine template and that they represent a novel
class of compounds that can be useful to design small molecules
with antiamyloid aggregation properties.
yield by increasing the base equivalence along with that of the
R-groups resulted in mixed isomers and/or double substitutions
as confirmed by NMR studies. Based on these observations,
selective alkylation at either N2 or N4-position was achieved
using K2CO3/DMA and NaH/DMSO, respectively, using
various substrates (Scheme 1, R = benzyl, substituted benzyl,
phenethyl, n-Pr, i-Pr, or cyclohexylmethyl). Attempts at
synthesizing the N2- or N4-methylpyridyl based DAQ
compounds using the same conditions were not successful.
The assessment of these isomeric DAQ derivatives against the
self-induced aggregation mechanisms of both Aβ40 and Aβ42
revealed a combination of structure−activity relationships
(SARs). As shown in Table 1, the aggregation IC50 values are
listed for each derivative and were compared with standard
controls (orange G, curcumin, resveratrol) and the DAQ
template itself. Examining the SAR data obtained based on a
thioflavin T (ThT) based fluorescence spectroscopy method
(Table 1) demonstrates the overall ability of these N2- and N4-
substituted DAQ derivatives to inhibit the aggregation of
amyloid peptides. In contrast, the DAQ template alone (Table
1) was promoting the aggregation process. Generally, N4-
substituted DAQ derivatives (3a−q) were more effective at
inhibiting Aβ40 compared to their N2-substituted isomers (4a−
q). Interestingly, this observation is completely reversed with
respect to Aβ42. The addition of a benzyl substituent either at
N4- or N2-position in compounds 3a and 4a provided Aβ40
aggregation inhibition (Table 1). However, the N4-regioisomer
3a did not exhibit any inhibition of Aβ42 unlike the N2-
regioisomer 4a. The N2-placement in 4a was more effective
against Aβ42 (IC50 = 8.4 μM). Replacing the benzyl substituent
with a more lipophilic 3- or 4-methylbenzyl substituent in 3b,
3c and 4b, 4c modified the biological profiles significantly.
When compared to 3a, the N4-placement of a methylbenzyl
substituent, regardless of meta- or para-positioning, slightly
improved activity against Aβ40 (no more than ∼1.3-fold
improvement) with no change toward Aβ42. This was not the
case with the N2-isomers, where the methyl-substituted benzyl
group resulted in loss of activity against both Aβ40 and Aβ42
(4b, Aβ40/42 IC50 > 25 μM) or a much weaker profile (4c,
Aβ40 IC50 = 13.1 μM, Aβ42 IC50 = 22.5 μM) compared to 4a
(Table 1). The addition of an electron-donating methoxybenzyl
group, in 3d, 3e and 4d, 4e, yielded mixed outcomes. The N4-
placement of a 3-methoxybenzyl group (3d, Aβ40 IC50 = 20.6
μM, Aβ42 IC50 = inactive at 25 μM) was detrimental to
antiaggregation activity compared to the benzyl derivative 3a,
while the 4-methoxybenzyl substituent (3e, Aβ40 IC50 = 1.1
μM, Aβ42 IC50 > 25 μM) enhanced the activity against Aβ40
by ∼4−4.5-fold (compared to 3a and 3c). However, N2-
placement of the 3-methoxy-substituted benzyl group (4d,
Aβ40/42 IC50 > 25 μM) did not offer any benefits compared to
4a and was ineffective in providing antiaggregation activity. The
The DAQ template (2, Scheme 1) was synthesized starting
from 2-fluoro (1a) or 2-aminobenzonitrile (1b) by heating with
a
Scheme 1
a
Reagents and conditions: (a) Alkyl- or aromatic halides, NaH,
DMSO, 0 °C to r.t. 14 h. (b) Alkyl- or aromatic halides, potassium
carbonate, DMA, r.t. to 85 °C, 5 h.
excess guanidine carbonate in dimethylacetamide (DMA)
under pressure (Scheme S1, Supporting Information).18,19
The reaction with 2-fluorbenzonitrile provided superior yield
(∼70%) of 2 compared to 2-aminobenzonitrile (∼40%). The
most challenging aspect of this synthesis was developing
selective alkylation methods to prepare either N2- or N4-
substituted quinazoline amines. After investigating a variety of
base, solvent, time, and temperature combinations (Table S1,
Supporting Information), two selective conditions were
identified. Using NaH and DMSO provided the N4-isomer
exclusively, whereas combination of potassium carbonate and
DMA favored the N2-isomer exclusively (Scheme 1). Their
4-methoxy compound (4e, Aβ40 IC50 = 6.8 μM, Aβ42 IC50
>
25 μM) was ∼6-fold less potent compared to its N4-isomer
(3e) toward Aβ40.
The effect of electron-withdrawing trifluoromethylbenzyl
substituent was investigated in compounds 3f, 3g and 4f, 4g.
Interestingly, its presence at N4-position in compound 3f (3-
CF3-benzyl) provided similar inhibition profile compared to 3b
(3-Me) with IC50 values around 3.6 μM for Aβ40 and over 20
μM for Aβ42. This was not the case with 3g (4-CF3), which
was ∼2-fold more potent toward Aβ40 compared to its
methylbenzyl bioisostere (3c). However, this modification did
not provide better inhibition of Aβ42. With N2-placement,
1
chemical structures were confirmed by H, 13C NMR, and 2D
COSY NMR studies. These studies show that the N4-NHs are
more acidic (δ 7.74 ppm) compared to the N2-NH protons (δ
6.53 ppm).
The conditions used for selective alkylation provided low
yield (14−31%). The TLC examinations showed that the
reactions did not go to completion. Attempts to increase the
B
DOI: 10.1021/acsmedchemlett.6b00039
ACS Med. Chem. Lett. XXXX, XXX, XXX−XXX

ACS Medicinal Chemistry Letters
Letter
Table 1. Inhibition Data for DAQ Isomeric Derivatives 3a−q and 4a−q against Self-Induced Aβ40 and Aβ42 Aggregation
ab
,
ab
,
IC50 (μM)
IC50 (μM)
N4-isomer
Aβ40
Aβ42
R-group
benzyl
Aβ40
Aβ42
N2-isomer
3a
4.8
>25
>25
>25
n.a
2.2
8.4
4a
3b
3.6
3-Me-benzyl
4-Me-benzyl
3-OMe-benzyl
4-OMe-benzyl
3-CF3-benzyl
4-CF3-benzyl
3-Cl-benzyl
4-Cl-benzyl
3-Br-benzyl
4-Br-benzyl
3-F-benzyl
4-F-benzyl
phenethyl
>25
13.1
>25
6.8
>25
22.5
>25
>25
13.8
5.3
4b
3c
3.9
4c
3d
20.6
1.1
4d
3e
>25
21.0
>25
>25
>25
22.3
14.8
>25
>25
>25
n.a
4e
3f
3.5
12.5
2.3
4f
3g
1.9
4g
3h
1.9
5.9
13.4
8.1
4h
3i
0.62
0.58
0.08
2.8
0.93
1.5
4i
3j
2.7
4j
3k
1.7
1.7
4k
3l
11.3
>25
14.9
>25
>25
>25
1.5
14.7
>25
>25
>25
>25
>25
3.1
4l
3m
3n
3.1
4m
7.8
4n
3o
14.9
10.2
13.6
P.A
3.0
n-Pr
4o
3p
n.a
i-Pr
4p
3q
>25
P.A
8.7
cyclohexylmethyl
4q
DAQ
orange G
curcumin
resveratrol
2.6
4.1
a
b
IC50 values are calculated based on the ThT-based fluorescence spectroscopy assay (excitation = 440 nm, emission = 490 nm). Values are mean
values of triplicate readings for three independent experiments. n.a = not active. P.A = promotes aggregation.
however, the modification from (3/4-Me) to (3/4-CF3) was
positive across the board. While 4b was ineffective against both
Aβ40/42 (IC50 > 25 μM), 4f (3-CF3) was active toward both
species (IC50 ≈ 12−13 μM). Derivative 4g, however, was more
potent compared to 4c (4-Me), equipotent to 4a (Aβ40), and
∼1.6-fold more potent toward Aβ42 compared to 4a. It
exhibited equipotent/comparable activity to the reference
compound resveratrol (Table 1). In the next set of compounds,
the effect of benzyl halides at N2 and N4-position was explored.
Generally, the presence of benzyl halides increased overall
inhibition potency (with the exception of fluorobenzyl).
Starting with the N4-placement of 3-chloro (3h) or 4-
chlorobenzyl (3i) substituents, both were more potent toward
Aβ40 compared to 3a (∼2.5-fold and ∼8-fold, respectively). In
this regard 3i exhibited potent inhibition of Aβ40 with an IC50
= 620 nM and was approximately 2.4-fold more potent
compared to the reference agent curcumin (Aβ IC50 = 1.5 μM).
and resveratrol. Furthermore, placing the 3-bromobenzyl
substituent at N2-position, significantly enhances the Aβ42
aggregation inhibition in compound 4j (Aβ40 IC50 = 2.7 μM)
with almost 8-fold gain in activity compared to the
corresponding N4-isomer 3j (Aβ42 IC50 = 22.3 μM, Table 1).
In contrast, the presence of an electronegative fluorobenzyl
substituent at either N2- or N4-position was detrimental
suggesting that smaller size of a fluorine atom compared to
either a bromine or chlorine atom was not favorable. The N4-
derivatives 3l and 3m were ineffective against Aβ42. Although
they lost considerable activity toward Aβ40, they did maintain
moderate levels of inhibition (IC50 ≈ 3 μM) comparable to the
methyl- (3b, 3c) and trifluoromethyl-based derivatives (3f, 3g).
The extent of activity loss was greater with N2-placements,
where derivative 4l exhibited weak activity toward both Aβ40/
42 (IC50 ≈ 11−15 μM) and 4m was ineffective against both
Aβ40/42 (Table 1). In other SAR modifications, the addition of
a methylene group (phenethyl vs benzyl) in compounds 3n and
4n, did not improve the Aβ inhibitory activity compared to 3a
and 4a (Table 1). Compound 3n was ∼1.6-fold less potent
toward Aβ40 compared to 3a and not very effective toward
Aβ42 either. This modification also caused a loss of Aβ42
inhibitory activity for 4n along with a ∼ 6.7-fold decrease in
potency toward Aβ40. Similarly, the addition of alkyl
substituents (n-Pr and i-Pr) either at N2- or N4-position, as
seen in compounds 3o, 3p and 4o, 4p resulted in either weak
or no inhibition of aggregation (Table 1). Furthermore, the
addition of a cyclic nonaromatic substituent such as a
cyclohexylmethyl group (compound 3q and 4q) was
detrimental to antiaggregation activity clearly displaying the
However, both 3i and 3h were ineffective against Aβ42 (IC50
>
25 μM). Interestingly, the N2-placement of these chlorobenzyl
groups (4h/4i) exhibit inhibition of both Aβ40 and Aβ42
aggregation. They were not as potent as 3i and 3h against Aβ40
(Table 1). The bromine-based bioisosteres (3j, 3k and 4j, 4k)
were all more potent against both Aβ40/42 compared to their
chlorine-based counterparts with one exception (4k on Aβ40,
Table 1). Derivative 3k (4-bromobenzyl) was identified as the
most potent Aβ40 aggregation inhibitor (IC50 ≈ 80 nM). It was
∼7.5-fold more potent compared to its 3-bromobenzyl isomer
(3j, Aβ40 IC50 = 580 nM, Aβ42 IC50 = 22.3 μM) and ∼21-fold
more potent compared to its N2-isomer (4k, Aβ40 and Aβ42
IC50 = 1.7 μM). Both 3k and 3j were much more potent
inhibitors compared to reference agents orange G, curcumin,
C
DOI: 10.1021/acsmedchemlett.6b00039
ACS Med. Chem. Lett. XXXX, XXX, XXX−XXX

ACS Medicinal Chemistry Letters
Letter
Figure 2. (a) Amyloid aggregation kinetics study. Panels A and B: 5 μM Aβ40 with varying concentrations (1, 5, or 25 μM) of DAQ derivatives 3j
and 3k, respectively. Panels C and D: 5 μM Aβ42 with varying concentrations (1, 5, or 25 μM) of DAQ derivatives 4j and 4k, respectively.
Aggregation kinetics were monitored by ThT-fluorescence spectroscopy (excitation = 440 nm, emission = 490 nm) for 24 h at 37 °C in phosphate
buffer at pH 7.4. Results are based on three independent experiments in triplicate measurements. (b) Amyloid morphology analysis using
transmission electron microscopy (TEM) after 24 h incubation at 37 °C in phosphate buffer at pH 7.4. Panel A, 25 μM Aβ40 control; Panel B, 25
μM Aβ40 with 25 μM of DAQ derivative 3k; Panel C, 25 μM Aβ42 control; Panel D, 25 μM Aβ42 with 25 μM of DAQ derivative 4k. Scale: black/
white bars represent 500 nm.
importance of having planar aromatic substituents to exhibit
antiaggregation properties.
approximately 12.5 h for Aβ40 control curve. In this regard,
both compounds 3j and 3k were able to prevent the nucleation-
dependent aggregation process significantly, with a relatively
weak growth phase seen toward the end of the 24 h aggregation
period. Significantly, both of these compounds were exhibiting
potent antiamyloid aggregation properties at all the tested
concentrations (1, 5, and 25 μM, Figure 2a, panels A and B).
The aggregation kinetics plot of the corresponding N2-isomers
4j and 4k (Figure S1, panels A and B, Supporting Information)
displayed concentration-dependent inhibition via improving
monomer stability and reducing the rate of aggregation;
however, the antiaggregation effect was not as drastic as
compounds 3j and 3k. Compound 4j at 25 μM was able to
delay the aggregation compared to Aβ control curve (t50 ≈ 17 h
vs 12.5 h for Aβ control with no test compound), whereas 4k at
25 μM was able to exhibit complete inhibition of Aβ40
aggregation. These studies demonstrate that at 25 μM, N4-
isomers, compounds 3j, 3k, and the N2-isomer 4k were capable
of preventing the formation of Aβ aggregates including dimers,
trimers, oligomers, protofibrils, and fibrils.
In summary, based on this SAR study, we identified a total of
nine DAQ based compounds (3e, 3g−k, and 4i−k) that
exhibited superior or equipotent anti-Aβ aggregation activity
profile compared to reference agents orange G, curcumin, and
resveratrol (Table 1). Compounds 4j and 4k met or surpassed
the Aβ42 activity seen with curcumin (the most potent among
the reference agents used), while a total of 9 derivatives (3e,
3g−k and 4i−k) closely met or surpassed the Aβ40 activity
shown by the same reference agents. Derivative 3k, N4-(4-
bromobenzyl)quinazoline-2,4-diamine, was identified as the
most potent Aβ40 aggregation inhibitor (IC50 ≈ 80 nM), while
its N2-isomer, N2-(4-bromobenzyl)quinazoline-2,4-iamine (4k),
was identified as the most potent and dual Aβ40/42
aggregation inhibitor (IC50 ≈ 1.7 μM).
To better understand the modes of inhibition exerted by 3k
and 4k (and their regioisomers 3j and 4j), we investigated the
full aggregation kinetic plots (Figure 2a, panels A−D, and
Figure S1, Supporting Information) derived from the ThT
based aggregation assay used to determine the IC50 values listed
in Table 1. The assessment of the kinetic plots provided greater
insight into the potential mode of action for this class of
compounds. The aggregation kinetic plots for compounds 3j
and 3k (Figure 2a, panels A and B) show that both of these
derivatives achieve their superior inhibitory activity by
stabilizing the monomeric Aβ40 structures. The half time
(t50) that defines the time at which the fluorescence intensity
reaches the midpoint between pre- and postaggregation
baseline growth phase in the absence of test compounds was
With respect to Aβ42, the dominance of the N2-isomers (4j
and 4k, Figure 2a, panels C and D) became evident when
compared to their N4-counterparts (3j and 3k, Figure S1,
panels C and D, Supporting Information). The half time t50 was
approximately 2 h for Aβ42 in the absence of test compounds.
In this regard, compounds 3j and 3k were able to delay the lag
phase to a great extent (t50 values of approximately 2.6 and 3.5
h, respectively, at 25 μM compared to Aβ42 control. In
contrast, the N4-isomers 4j and 4k exhibited better inhibition
profile with compound 4j showing a significant delay in
D
DOI: 10.1021/acsmedchemlett.6b00039
ACS Med. Chem. Lett. XXXX, XXX, XXX−XXX

ACS Medicinal Chemistry Letters
Letter
aggregation (t50 ≈ 6 h), whereas 4k was able to show complete
inhibition at 25 μM (Figure 2a, panels C and D). While all the
kinetic plots for Aβ42 demonstrated concentration-dependent
inhibition, none of these derivatives were capable of completely
halting the saturation phase except for the N2-isomer 4k at the
highest concentration tested.
To further validate these observations, Aβ40/42 morphology
was assessed using transmission electron microscopy (TEM,
Figure 2b, Figure S2 and S3, Supporting Information) in the
presence and absence of most potent DAQ derivatives (3j, 3k,
4j, and 4k, Figure 2b and Figure S2, Supporting Information).
As observed with TEM, both Aβ40/42 control samples (25
μM) displayed high levels of dense and full aggregate species.
Coincubation experiments with DAQ derivatives (3j, 3k, 4j and
4k), at 25 μM, were able to drastically hinder the aggregation
process with the N4-isomer 3k (Figure 2b, panel B) and N2-
isomer 4k (Figure 2b, panel D) exhibiting almost complete
inhibition of aggregation. These results correlate with the ThT
aggregation kinetics data (Figure 2a and Figure S1, Supporting
Information).
In order to gain some insights into the mechanism of anti-Aβ
aggregation properties of DAQ derivatives (3j, 3k, 4j, and 4k),
we investigated their binding interactions in both the Aβ dimer
and Aβ fibril model (Figures S4 and S5, Supporting
Information). In the dimer model (Figure S4, panels A and
B, Supporting Information), featuring the Aβ9−40 fragment,20
each of the docked DAQ derivatives interacted with the dimeric
structure in a unique fashion, although these interactions were
generally localized in the dimer-turn region (Asp23-Gly29).
This might explain the trends seen in the kinetic plots, where
monomer stability and substantial delays in the aggregation
process were observed. The DAQ ring scaffold of 3j was
oriented between Phe20 (perpendicular to the phenyl ring) and
Val24 (distance ∼5−7 Å), where the 3-bromophenyl ring was
aligned, in a parallel fashion, toward Ser26 and Asn27 (distance
≈ 5 Å). The bromine atom was oriented toward Ile32-Leu34
pocket at the C-terminal end (Figure S4, panel A). The C2-
amine was positioned across from Val24-Gly25 on the adjacent
monomer (distance ≈ 5 Å). However, the corresponding N2-
isomer 4j had its DAQ ring scaffold between Asp23 and Lys28
(distance ≈ 5 Å), where the C4-amine was undergoing
hydrogen-bonding interaction with Asp23 (distance < 3 Å).
The 3-bromophenyl ring was oriented toward Asp23 and Val24
(distance ≈ 5 Å), where the C2−NH was undergoing
hydrogen-bonding with the Asp23 side-chain and backbone
carbonyl of Val24 (distance < 3 Å). The 4-bromobenzyl
isomers (3k and 4k) exhibited nearly opposite binding modes
in the dimer model (Figure S4, panel B). The DAQ ring
scaffold of 3k was oriented parallel to the Ser26-Lys28 turn
region, where the C2-amine was in contact with the Asp23 side-
chain and backbone carbonyl of Val24 (distance = 2.7−3.1 Å).
The 4-bromophenyl group was between Asp23 and Ile31
(distance ≈ 5−6 Å). In contrast, the corresponding N2-isomer
4k had its DAQ ring scaffold oriented between Ile31/32 and
Ala21-Asp23 in a perpendicular fashion, where the C4-amine
was hydrogen-bonding with the carbonyl backbone of Ile32
(distance ≈ 2.7 Å). The C2−NH was oriented toward Ala30
and was in contact with its carbonyl backbone (distance ≈ 3 Å),
while the 4-bromophenyl group was stacked, in a parallel
orientation, between Asp23 and Gly29 (distance ≈ 5 Å). These
studies suggest that DAQ ring system serves as a suitable
template to design small molecule probes to study Aβ
aggregation and inhibition.
In conclusion, we investigated the selective alkylation of the
2,4-diaminoquinazoline (DAQ) template, a privileged scaffold,
to generate a library of N2 and N4-substituted DAQ derivatives.
These compounds were then screened for antiaggregation
properties toward Aβ40/42 by monitoring their aggregation
kinetics, which revealed that halogen-substituted benzyl groups
generally exhibited superior anti-Aβ aggregation effect with N4-
isomers providing better selectivity for Aβ40, whereas the N2-
isomers exhibited better inhibition of Aβ42 aggregation. The
N4-isomer 3k with a 4-bromobenzyl substituent was identified
as the most potent Aβ40 aggregation inhibitor (IC50 = 80 nM),
whereas the corresponding N2-isomer (4k) yielded our most
potent Aβ42 aggregation inhibitor (IC50 = 1.7 μM), which also
exhibited dual Aβ40/42 aggregation inhibition. The outcomes
of this study demonstrates the usefulness of quinazoline
diamine template to design novel antiamyloid agents. These
small molecules serve as valuable pharmacological tools to
study and develop potential therapies to treat AD.
ASSOCIATED CONTENT
* Supporting Information
■
S
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acsmedchem-
lett.6b00039.
Synthetic and biological methods along with character-
ization and analytical data (PDF)
AUTHOR INFORMATION
Corresponding Author
*Tel: 519-888-4567, ext. 21317. E-mail: praopera@uwaterloo.
■
ca.
Author Contributions
P.P.N.R. and T.M. conceived the project and designed the
experiments. T.M., A.S., and G.T., performed the experiments.
T.M., A.S., and P.P.N.R. analyzed and interpreted the data.
T.M. wrote the manuscript. T.M., A.S., G.T., and P.P.N.R.
revised the manuscript.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors would like to thank the Faculty of Science, Office
of Research, the School of Pharmacy at the University of
Waterloo, Ontario Mental Health Foundation (graduate
scholarship for T.M.), NSERC-Discovery (RGPIN: 03830-
2014), Canada Foundation for Innovation (CFI-JELF), Ontario
Research Fund (ORF), and Early Researcher Award, Ministry
of Research and Innovation, Government of Ontario, Canada
(PR) for financial support of this research project.
ABBREVIATIONS
■
AD, Alzheimer’s disease; Aβ, amyloid-beta; DAQ, diaminoqui-
nazoline; DMA, dimethylacetamide; DMSO, dimethyl sulf-
oxide; NaH, sodium hydride; SAR, structure−activity relation-
ship; DMAP, 4-dimethylaminopyridine; DBU, 1,8-diazabicy-
cloundec-7-ene
REFERENCES
■
(1) Selkoe, D. J. Cell biology of the amyloid β protein precursor and
the mechanism of Alzheimer’s disease. Annu. Rev. Cell Biol. 1994, 10,
373−403.
E
DOI: 10.1021/acsmedchemlett.6b00039
ACS Med. Chem. Lett. XXXX, XXX, XXX−XXX

ACS Medicinal Chemistry Letters
Letter
(2) LaFerla, F. M.; Green, K. N.; Oddo, S. Intracellular amyloid-β in
Alzheimer’s disease. Nat. Rev. Neurosci. 2007, 8, 499−509.
(3) Price, D. L.; Sisodia, S. S.; Gandy, S. E. Amyloid beta amyloidosis
in Alzheimer’s disease. Curr. Opin. Neurol. 1995, 8, 268−274.
(4) Guela, C.; Wu, C.; Saroff, D.; Lorenzo, A.; Yuan, M.; Yanker, B.
A. Aging renders the brain vulnerable to amyloid β-protein
neurotoxicity. Nat. Med. 1998, 4, 827−831.
(5) Hock, C.; Konietzko, U.; Streffer, J. R.; Tracy, J.; Signorell, A.;
Muller-Tillmanns, B.; Lemke, U.; Henke, K.; Moritz, E.; Garcia, E.;
Wollmer, M. A.; Umbricht, D.; de Quervain, D. J. F.; Hofmann, M.;
Maddalena, A.; Papassotiropoulos, A.; Nitsch, R. M. Antibodies against
β-amyloid slow cognitive decline in Alzheimer’s disease. Neuron 2003,
38, 547−554.
(6) Menendez-Gonzalez, M. Biomarkers in neurodegenerative
disorders: translating research into clinical practice. Front. Aging
Neurosci. 2014, 6, 1−2.
(7) Godyn, J.; Jonczyk, J.; Panek, D.; Malawska, B. Therapeutic
strategies for Alzheimer’s disease in clinical trials. Pharmacol. Rep.
2016, 68, 127−138.
(8) Masliah, E.; Rockenstein, E.; Price, D.; Bonhaus, D.; Meier, D.;
Moesler, H.; Konrat, R.; Tsigelny, I.; Wrasidlo, W. Small molecules
targeting the formation of neurotoxic oligomers for the treatment of
AD/PD: a second look. Neurobiol. Aging 2014, 35, S15.
(9) Doig, A. J.; Derreumaux, P. Inhibition of protein aggregation and
amyloid formation by small molecules. Curr. Opin. Struct. Biol. 2015,
30, 50−56.
(10) Ono, K.; Hasegawa, K.; Naiki, H.; Yamanda, M. Curcumin has
potent anti-amyloidogenic effects for Alzheimer’s β-amyloid fibrils in
vitro. J. Neurosci. Res. 2004, 75, 742−750.
(11) Hu, S.; Maiti, P.; Ma, Q.; Zuo, X.; Jones, M. R.; Cole, G. M.;
Frautschy, S. A. Clinical development of curcumin in neuro-
degenerative disease. Expert Rev. Neurother. 2015, 15, 629−637.
(12) Granzotto, A.; Zatta, P. Resveratrol and Alzheimer’s disease:
message in a bottle on red wine and cognition. Front. Aging Neurosci.
2014, 6, 95−102.
(13) Bastianetto, S.; Menard, C.; Quirion, R. Neuroprotective action
of resveratrol. Biochim. Biophys. Acta, Mol. Basis Dis. 2015, 1852,
1195−1201.
(14) Jiang, L.; Liu, C.; Leibly, D.; Landau, M.; Zhao, M.; Hughes, M.
P.; Eisenberg, D. S. Structure-based discovery of fiber-binding
compounds that reduce the cytotoxicity of amyloid beta. eLife 2013,
2, e00857.
(15) Rao, P. P. N.; Mohamed, T.; Teckwani, K.; Tin, G. Curcumin
binding to beta amyloid: A computational study. Chem. Biol. Drug Des.
2015, 86, 813−820.
(16) Hamley, I. W. The Amyloid beta peptide: A chemist’s
perspective. Role in Alzheimer’s and fibrillization. Chem. Rev. 2012,
112, 5147−5192.
(17) Soto, C.; Kindy, M. S.; Baumann, M.; Frangione, B. Inhibition of
Alzhemier’s amyloidosis by peptides that prevent β-sheet conforma-
tion. Biochem. Biophys. Res. Commun. 1996, 226, 672−680.
(18) Chao, B.; Tong, X.; Tang, W.; Li, D.; He, P.; Garcia, J.; Zeng, L.;
Gao, A.; Yang, L.; Li, J.; Nan, F.; Jacobs, M.; Altmeyer, R.; Zuo, J.; Hu,
Y. Discovery and optimization of 2,4-diaminoquinazoline derivatives as
a new class of potent Dengue virus inhibitors. J. Med. Chem. 2012, 55,
3135−3143.
(19) Thurmond, J.; Butchbach, M. E. R.; Palomo, M.; Pease, B.; Rao,
M.; Bedell, L.; Keyvan, M.; Pai, G.; Mishra, R.; Haraldsson, M.;
Andresson, T.; Bragason, G.; Thosteinsdottir, M.; Bjornsson, J. M.;
Coovert, D. D.; Burghes, A. H. M.; Gurney, M. E.; Singh, J. Synthesis
and biological evaluation of novel 2,4-diaminoquinazoline derivatives
as SMN2 promoter activators for the potential treatment of spinal
muscular atrophy. J. Med. Chem. 2008, 51, 449−469.
(20) Petkova, A. T.; Yau, W. M.; Tycko, R. Experimental constraints
on quaternary structure in Alzheimer’s β-amyloid fibrils. Biochemistry
2006, 45, 498−512.
F
DOI: 10.1021/acsmedchemlett.6b00039
ACS Med. Chem. Lett. XXXX, XXX, XXX−XXX