Discovery and mechanism study of SIRT1 activators that promote the deacetylation of fluorophore-labeled substrate

Article abstract of DOI:10.1021/jm301032j

SIRT1 is an NAD⁺-dependent deacetylase, whose activators have potential therapeutic applications in age-related diseases. Here we report a new class of SIRT1 activators. The activation is dependent on the fluorophore labeled to the substrate. To elucidate the activation mechanism, we solved the crystal structure of SIRT3/ac-RHKK_ac-AMC complex. The structure revealed that the fluorophore blocked the H-bond formation and created a cavity between the substrate and the Rossmann fold. We built the SIRT1/ac-RHKK _ac-AMC complex model based on the crystal structure. K_m and K_d determinations demonstrated that the fluorophore decreased the peptide binding affinity. The binding modes of SIRT1 activators indicated that a portion of the activators interacts with the fluorophore through π-stacking, while the other portion inserts into the cavity or interacts with the Rossmann fold, thus increasing the substrate affinity. Our study provides new insights into the mechanism of SIRT1 activation and may aid the design of novel SIRT1 activators.

Full text of DOI:10.1021/jm301032j

Article
pubs.acs.org/jmc
Discovery and Mechanism Study of SIRT1 Activators that Promote
the Deacetylation of Fluorophore-Labeled Substrate
Jiahui Wu,†,▽ Dengyou Zhang,‡,▽ Lei Chen,† Jianneng Li,⊥ Jianling Wang,§ Chengqing Ning,#
Niefang Yu,# Fei Zhao,‡ Dongying Chen,§ Xiaoyan Chen,∥ Kaixian Chen,‡ Hualiang Jiang,‡ Hong Liu,*,‡
and Dongxiang Liu*,†,¶
†Department of Pharmacology III, ‡Drug Discovery and Design Center, §Department of Analytical Chemistry, and ∥Center for Drug
Metabolism and Pharmacokinetics, Shanghai Institute of Materia Medica, Chinese Academy of Sciences, 555 Zu-Chong-Zhi Road,
Shanghai 201203, China
⊥School of Life Sciences, Fudan University, Shanghai 200433, China
#Institute of Molecular Design and Drug Discovery, School of Pharmaceutical Sciences, Central South University, Changsha 410078,
Hunan, China
S
* Supporting Information
ABSTRACT: SIRT1 is an NAD+-dependent deacetylase, whose activators have
potential therapeutic applications in age-related diseases. Here we report a new
class of SIRT1 activators. The activation is dependent on the fluorophore labeled
to the substrate. To elucidate the activation mechanism, we solved the crystal
structure of SIRT3/ac-RHKKac-AMC complex. The structure revealed that the
fluorophore blocked the H-bond formation and created a cavity between the
substrate and the Rossmann fold. We built the SIRT1/ac-RHKKac-AMC complex
model based on the crystal structure. Km and Kd determinations demonstrated
that the fluorophore decreased the peptide binding affinity. The binding modes of
SIRT1 activators indicated that a portion of the activators interacts with the fluorophore through π-stacking, while the other
portion inserts into the cavity or interacts with the Rossmann fold, thus increasing the substrate affinity. Our study provides new
insights into the mechanism of SIRT1 activation and may aid the design of novel SIRT1 activators.
factors.7 Besides, SIRT1 can regulate the genes involved in
insulin secretion in pancreatic β cell lines and β cell-specific
INTRODUCTION
■
Sirtuins are a class of protein deacetylases that belong to KDAC
SIRT1-overexpressing (BESTO) mice by repressing uncou-
family. Unlike other KDACs, the deacetylation by sirtuins
pling protein 2 (Ucp2).8,9 Additionally, SIRT1 promotes
requires the involvement of nicotinamide adenine dinucleotide
(NAD+).1 Mammal possesses seven sirtuins (SIRT1−SIRT7),
glucose homeostasis and mitochondrial biogenesis through
PPARγ coactivator 1α (PGC-1α).10,11 Therefore, SIRT1 is
and they are widely distributed in different subcellular
compartments.2 SIRT1, SIRT6, and SIRT7 are localized in
involved in a wide range of physiological activities including
fatty acid oxidization, stress resistance, insulin secretion, and
the nucleus, while SIRT2 is in the cytoplasm and SIRT3,
glucose production, which are closely related to age-related
SIRT4, and SIRT5 are in mitochondria. Among them, SIRT1
diseases such as type 2 diabetes, cardiovascular disease,
shares the highest sequence homology with yeast sirtuin
metabolic syndrome, inflammation, and also to aging and
information regulator 2 (Sir2) and is the most extensively
studied.
longevity.12,13 Development of SIRT1 activators has potential
therapeutic applications in the treatment of the age-related
SIRT1 deacetylates a broad range of substrates and plays
diseases.
important roles in many biological functions. For instance,
Up to the present, several SIRT1 activators were identified
(Figure 1). They could be mainly categorized into two classes:
natural plant polyphenols and synthesized small molecules. 1
(resveratrol) and several other plant polyphenols were first
SIRT1 deacetylates histone H1 and promotes the hetero-
chromatin formation.3 H3 and H4 could be deacetylated by
SIRT1 and the RNA interference of SIRT1 induced hyper-
acetylation of the two proteins in human cells.4 SIRT1 binds to
discovered by using a fluorescent assay with a fluorophore-
the fat regulator peroxisome proliferator activated receptor γ
labeled peptide as the substrate.14 Treatment of Saccharomyces
(PPARγ) and represses the genes controlled by PPARγ. Studies
cerevisiae with 1 significantly increased the life span of the yeast.
revealed that overexpression of SIRT1 attenuates the adipo-
Studies in Caenorhabditis elegans,15 Drosophila,15,16 and
genesis.5 Also, adiponectin transcription could be enhanced by
SIRT1 in adipocytes through activating FOXO16 and cell
resistance to oxidative stress could be increased by SIRT1
through regulating FOXO family of Forkhead transcription
Received: July 13, 2012
Published: January 14, 2013
© 2013 American Chemical Society
761
dx.doi.org/10.1021/jm301032j | J. Med. Chem. 2013, 56, 761−780

Journal of Medicinal Chemistry
Article
Figure 1. Chemical structures of representative SIRT1 activators.
vertebrates17 had proven the positive effect of 1 in increasing
the life span. Besides, 1 was found to attenuated insulin
resistance under high-fat diet induced conditions in vivo,18 and
treatment of type 2 diabetic mice with 1 induced mitochondrial
biogenesis.19 Meanwhile, small molecules with structures
unrelated to 1 had been identified to activate SIRT1 in vitro
by using the fluorophore-labeled substrates, which include
compound 2,20 3,21 and the compounds discovered by the
Sirtris company like 4 (SRT1460), 5 (SRT1720),22,23 and 624
(Figure 1).
However, it is still controversial whether 1 and Sirtris
compounds are real SIRT1 activators. In vitro studies revealed
that the activation of SIRT1 by these compounds requires the
fluorophore attached to the peptide substrate, and the
activation could not be observed in deacetylating the substrate
lacking of fluorophore.25,26 Even though Dai et al. provided the
biophysical evidence of the direct binding between the Sirtris
compounds and SIRT1, they suggested that the activators bind
in an allosteric site of the enzyme.27 In terms of in vivo effect of
1, in contrast with the positive results mentioned above, some
studies showed that there is neither detectable effect of 1 on
Sir2 activity in yeast25 nor significant effect on lifespan in
Drosophila.28 As for the SIRT1 small molecule activator 5, some
results indicated that it could ameliorate insulin sensitivity and
glucose tolerance in obese mice22,29 and enhance mitochondrial
biogenesis in mice fed with a high calorie diet,30 whereas
Pacholec et al. reported that no effect of lowering plasma
glucose or improving mitochondrial capacity could be detected
in mice.26
Considering the numerous and paradoxical reports about
SIRT1 activators, obviously it is very important to elucidate the
activation mechanism of the activators on molecular level. In
this study, a series of diaryl acylhydrazone compounds with the
activation activities of SIRT1 were first identified from our in-
house compound library by employing a coumarin-labeled
acetyl peptide. We synthesized a series of analogues of these
compounds and established their structure−activity relation-
ship. Combining the protein crystallography, molecular
modeling, and site-directed protein mutagenesis, we deter-
mined the binding site of our compounds, 1 and 4. The activity,
stability, and bioavailability of our compounds were also tested
on cell level. Our goal is to explain the substrate selectivity and
the activation mechanism of these SIRT1 activators from the
structural point of view. The mechanism of SIRT1 activation by
small molecules presented here may shed light in the design of
SIRT1 novel activators.
RESULTS AND DISCUSSION
■
Identification of Chemical Compounds That Enhance
SIRT1 Enzyme Activity. A high-throughput fluorescent assay
was deployed to screen our in-house compound library aiming
to discover novel SIRT1 activators. To our delight, five
compounds based on diaryl acylhydrazone scaffold were
identified with the activation effect of SIRT1 deacetylation
(Table 1). Notably, our compounds displayed EC1.5 values
ranging from 1.6 to 7.0 μM and the highest maximum
activation was up to 375%. The EC1.5 values were better than
that of 1 (EC1.5 of 1 was 23.6 μM). These results encouraged us
to further optimize the diaryl acylhydrazone scaffold. Our
efforts were mainly focused on diversifying the aryl fragments
due to their convenient modification.
Synthesis of Diaryl Acylhydrozone Analogues. Com-
pounds 16a−n were prepared according to Scheme 1. Suzuki
coupling of methyl 4-bromobenzoate 12a with aryl bromide or
4-(methoxycarbonyl)phenylboronic acid 12b with aryl boric
acid smoothly afforded biphenyl esters 13a−k. Hydrazinolyza-
tion of 13a−k and 12c gave hydrozide intermediate 14a−k.
Condensation of 14a−k with benzaldehyde moieties 15a−d
readily provided compounds 16a−k. 15b−d were prepared
according to the refs 31 and 32. Preparation of compounds
16l−n was from commercial available hydrazides 17l−n.
Naphthyl analogues 20a−i were synthesized in a similar
manner as 16a−n (Scheme 2). The details of the synthetic
procedures and structural characterizations are described in the
Experimental Section.
Structure−Activity Relationship (SAR) of Diaryl
Acylhydrazone Analogues. The activation curves of
representative compounds were shown in Figure 2. Initial
observation of the activities (Table 1) revealed that the
compound bearing the p-Cl phenyl group (9) seemed to be
inferior to analogues with the biphenyl (7, 8) or naphthyl
762
dx.doi.org/10.1021/jm301032j | J. Med. Chem. 2013, 56, 761−780

Journal of Medicinal Chemistry
Article
c
Table 1. The Activities of the Initial Hits Identified by Our Fluorescent Assay
a
b
EC1.5 value represents the concentration of an activator required to increase the enzyme activity by 50%. Percentage of maximum activation
c
achieved. For comparison, the activity of compound 1 was also listed in this table.
a
Scheme 1. Synthesis of Compounds 16a−n
a
Reagents and conditions: (i) Pd(PPh3)4, KF·2H2O, DME/H2O/EtOH (7:3:2), Mw, 110 °C; (ii) N2H4·H2O, MeOH, reflux; (iii) AcOH, EtOH,
reflux.
763
dx.doi.org/10.1021/jm301032j | J. Med. Chem. 2013, 56, 761−780

Journal of Medicinal Chemistry
Article
a
Scheme 2. Synthesis of Compounds 20a−i
a
Reagents and conditions: (i) Pd(PPh3)4, KF·2H2O, DME/H2O/EtOH (7:3:2), Mw, 110 °C; (ii) N2H4·H2O, MeOH, reflux; (iii) AcOH, EtOH,
reflux.
(EC1.5 = 55.2 μM) resulted in a sharp decrease of the activity.
Introduction of the bromide (20b) at the C-6 position of the
naphthalene (P2 part) displayed a similar EC1.5 (2.7 μM) but
with the improvement of maximum activation up to 647%. A
block of either hydroxyl at the C-2 and C-4 positions (P1 part)
using methyl group (20c and 20d) was not beneficial for the
potency, which was also observed in biphenyl analogues.
Changing the bromide to electron-withdrawing groups such as
a methylsulfonyl (20e) or a cyano group (20f) gave rise to a
decrease of maximum activation. Appending the methoxy group
(20g) increased the EC1.5 to 0.9 μM and the maximum
activation to 624%, which showed the highest activation
potency among the compounds. Although the incorporation of
furan (20h) and benzene (20i) at the C-6 position of the
naphthalene (P2 part) maintained a comparable EC1.5, the
maximum activation tended to decrease.
To examine the selectivity of the SIRT1 activators, we tested
their activities on SIRT1 and its closest sirtuin homologues:
SIRT2 and SIRT3. The compounds were tested at the
concentration of 50 μM. The data for SIRT1, SIRT2, and
Figure 2. The dose−response curves of our compounds and 1
measured by the fluorescent assay.
SIRT3 were listed in Supporting Information Table S1. The
EC1.5 values for SIRT2 and SIRT3 of the compounds
unmentioned in the following parentheses were greater than
50 μM (SIRT2: 8, EC1.5 = 16.5 μM, maximum activation =
206%. 20i, EC1.5 = 11.6 μM, maximum activation = 257%.
SIRT3: 20i, EC1.5 = 28.6 μM, maximum activation = 199%.).
The results showed that these compounds (1, 7−11, 16a−n,
20a−i) are selective for the activation of SIRT1.
groups (10, 11). Changing the p-Cl phenyl group to other
phenyl groups such as p-tolyl (16l) and p-CF3 phenyl (16m)
groups slightly reduced the activity (Table 2). Introduction of
3,5-dimethoxyl phenyl (16n) group notably weakened SIRT1
activation. These results suggested that the modification of
phenyl group might not give rise to remarkable gain of the
activity, which turned our attention to optimize biphenyl and
naphthyl analogues.
Compared with the parent compound 7 (EC1.5 = 1.7 μM,
maximum activation = 246%), introduction of various
substituted phenyl moieties 16a−c at the C-4 position (P2
Part) retained the comparable activity whereas the C-3
substituted analogue 16d exhibited 10-fold less activity (EC1.5
= 18.3 μM), indicating that a linear fragment was preferred over
a branched fragment. 3,5-Dibromide substituents (P1 part)
were shown to play an important role in binding affinity
because the removal of two bromide 16e resulted in a drop-off
of EC1.5 (10.4 μM). Methylation of either hydroxy group (16f
and 16g) also eroded the activity, which indicated that both
hydroxyl groups at C-2 and C-4 positions (P1 part) might
function as hydrogen donors. Replacement of the phenyl
moieties with heteroaryl groups (P2 part) such as furan 16h
and pyridine 16i−k maintained the similar EC1.5 but with the
improvement of the maximum activation.
Compounds Promote the Deacetylation of Fluoro-
phore-Labeled Peptide but Not the Unlabeled Peptide.
Further, we tested the activities of 20b, 1, and 4 on
fluorophore-unlabeled acetyl peptide. A known SIRT1
inhibitor, 6-chloro-2,3,4,9-tetrahydro-1H-carbazole-1-carboxa-
mide (Ex527), was used as control.33 Acetyl-Arg-His-Lys-Lys
(ε-acetyl)-NH2 peptide (ac-RHKKac-NH2) was synthesized
without 4-amino-7-methylcoumarin (AMC) labeled at the C-
terminus. The deacetylation reactions of the unlabeled peptide
were carried out in the same way as the AMC-labeled peptide.
The amount of deacetylated peptide was determined by HPLC.
Deacetylation enhancement by compound 20b was observed
with the AMC-labeled peptide but not the unlabeled peptide
(Supporting Information Figure S3 and Figure 3). Inconsistent
with previous reports,25,26 1 and 4 did not activate SIRT1 in
deacetylating the fluorophore-unlabeled peptide. This phenom-
enon aroused our guess that the compounds do not bind
directly with SIRT1 but bind to the “SIRT1/fluorophore-
labeled peptide” complex. ITC experiments carried out by
Milne et al. and Pacholec et al. also support the conjecture.22,26
Next, a series of naphthyl analogues were explored (Table 3).
Interestingly, replacement of the naphthalene 11 (EC1.5 = 1.6
μM, maximum activation = 331%) with the quinoline 20a
764
dx.doi.org/10.1021/jm301032j | J. Med. Chem. 2013, 56, 761−780

Journal of Medicinal Chemistry
Article
d
Table 2. Structures and Activities of the Biphenyl and Phenyl Hydrazones
a
b
EC1.5 value represents the concentration of an activator required to increase the enzyme activity by 50%. Percentage of maximum activation
c
d
achieved. None detected. * = R1 at C-3 position; # = R1 at C-3 and C-5 position.
When 4 was titrated to SIRT1-substrate (Tamra-labeled
peptide) mixture, the binding was observed.22,26 However, no
binding isotherm was observed when 4 was titrated to SIRT1
and fluorophore-unlabeled peptide solution.26 Thus, we believe
that the fluorophore linked to the peptide is the key element for
the binding activity.
Crystal Structure of ac-RHKKac-AMC Peptide in
Complex with SIRT3. To confirm the role of the fluorophore,
we tried to determine the conformation of the fluorophore
conjugated with the peptide inside the enzyme. Up to now, the
structures of several sirtuin proteins were solved, including
Hst2 and Sir2p of yeast, Sir2-Af1 and Sir2-Af2 of archaea,
Sir2Tm of bacteria, and SIRT2, SIRT3, SIRT6, and SIRT7 of
humans. However, the structure of SIRT1 is still uncovered due
to the aggregation feature of the protein. Both sequence
alignment and structure superposition of the sirtuin homo-
logues revealed that they share high sequence homology as well
as structure homology.34 This suggests that SIRT1 model may
765
dx.doi.org/10.1021/jm301032j | J. Med. Chem. 2013, 56, 761−780

Journal of Medicinal Chemistry
Article
Table 3. Structures and Activities of the Naphthohydrazones
a
EC1.5 value represents the concentration of an activator required to
b
increase the enzyme activity by 50%. Percentage of maximum
activation achieved. None detected.
c
Table 4. Data Collection and Refinement Statistics of
“SIRT3/ac-RHKKac-AMC” Complex
diffraction data
space group
R32
a
resolution (Å)
50−2.1 (2.18−2.10)
unit cell parameters
a, b, c (Å)
114.809, 114.809, 123.202
90.00, 90.00, 120.00
99.9 (100.0)
α, β, γ (deg)
Figure 3. The Lineweaver−Burk plots of (A) ac-RHKKac-AMC
peptide and (B) ac-RHKKac-NH2 peptide in the presence of 50 μM 4
and 10 μM 20b, which were determined by HPLC method. 4 and 20b
exhibited substrate selectivity against native peptide and fluorophore-
labeled peptide.
a
completeness (%)
a
redundancy
11.0 (11.1)
a
average I/σI
24.8 (4.4)
a
Rmerge (%)
10.0 (66.5)
refinement statistics
Rworking (%)
20.76
peptide binding site, the catalytic core domain of SIRT1 could
be built by using the structure of SIRT3 as template. Thus,
SIRT3 was chosen for cocrystallization with the AMC-labeled
acetyl peptide.
Rfree (%)
28.19
rmsd in bond length (Å)
rmsd in bond angles (deg)
mean B-factors
0.0242
2.0687
45.762
The purified SIRT3 protein was incubated with ac-RHKKac-
AMC peptide at 4 °C overnight. The mixture was screened
with Hampton Research kits by using the hanging-drop method
at 15 °C. After two weeks, crystals were obtained with
crystallization solution containing 0.085 M Hepes-Na (pH 7.5),
8.5% (v/v) isopropyl alcohol, 17% (w/v) polyethylene glycol
(PEG) 4000, and 15% (v/v) glycerol anhydrous. The X-ray
diffractions of the crystals were collected at Shanghai
Synchrotron Radiation Facility. The data was processed by
HKL200035 and ccp4 suite.36 The data collection and
refinement statistics of the complex structure were listed in
Table 4. The space group of the complex is R32, which has one
molecule in an asymmetric unit. Interestingly, the AMC-labeled
peptides are involved in a “head-to-head” packing, which makes
the complexes exhibiting a dimeric arrangement (Figure 4A,B).
The acetyl-lysine (ALY) of the substrate inserts into the cleft
between the zinc-binding motif (small domain) and the
Rossmann fold region (large domain). The coumarins of the
two peptide substrates are in sandwich stacking, with the side
Ramachandran plot
in preferred regions (%)
in allowed regions (%)
outliers (%)
93.85
6.15
0.00
a
Values for the highest resolution shells are represented in
parentheses.
be built based on its homologue protein. According to the
sequence alignment of SIRT1 and SIRT3 (Supporting
Information Figure S4), the two proteins showed very high
sequence homology in the catalytic region. The differences
between SIRT1 and SIRT3 sequences were at the flexible loop
connecting the zinc-binding motif and Rossmann fold region
(between α2 and α3), a turn in the zinc-binding motif
(between α9 and β6), and a short α-helix on the Rossmann fold
domain (α12). These amino acids are distributed mainly on the
surface of SIRT3 protein, which are far away from the acetyl
peptide binding site. Because our attention was focused on the
766
dx.doi.org/10.1021/jm301032j | J. Med. Chem. 2013, 56, 761−780

Journal of Medicinal Chemistry
Article
Figure 4. Structure of human SIRT3 in complex with substrate ac-RHKKac-AMC. (A) Ribbon diagram of the overall structure of the complexes in
dimer. One molecule of SIRT3 protein is shown in green and the other in cyan. The substrate is shown in a stick representation; carbon, oxygen, and
nitrogen are colored in magentas, red, and blue, respectively. Zinc ions are highlighted in red. (B) An enlarged view of the binding site in dimeric
arrangement. The coumarin fluorophores of the substrates are packed in a “head-in-head” conformation, and the two peptides are represented with
the 2Fo − Fc electron density maps (1.0 σ-contoured). The labeled amino acids are involved in the packing. (C) The detailed interaction of Figure
4B without map. The black dashed lines represent the H-bonds formed between the substrate and the residues. (D) An enlarged view of the active
site in an asymmetric cell. Four hydrogen bonds (represented with black dashed lines) are formed between the substrate and the protein. Water
molecules are represented with yellow spheres, and the H-bond interactions of waters are shown as yellow dashed lines. Loop A and loop B were
indicated with arrows.
chain of arginine from another substrate packing aside the
fluorophore. Because of such substrate arrangement, the two
protein molecules are packed in a little twisted “shoulder-in-
shoulder” conformation. Moreover, the π-stacking was also
observed in the side chains of Trp 353 of two protein molecules
(Figure 4A). Though the dimerization of the fluorophores was
observed under the crystallization condition, the complex
structure suggests that the coumarin may intact with other
aromatic rings such as small molecules under physiological
conditions in like manner. This is because the configuration of a
single aromatic fluorophore is energetically unfavorable at
physiological conditions due to its high hydrophobicity and
lacking of significant contacts with the protein. The interaction
with another aromatic ring may stabilize the configuration of
the fluorophore.
The interaction details around the coumarins of two complex
molecules were shown in Figure 4C. As described above, the
arginine of one substrate packs head-to-head with the coumarin
from another substrate. The γ-guanido group of arginine is
usually protonated at physiological conditions. It may form
cation−π interaction with the coumarin and Phe 294. Also, the
arginine forms an H-bond with the carboxyl of Glu 181 of
another protein molecule. The alkyl side chain of the arginine
and the N-terminal acetyl group of the peptide are in van der
Waals interactions with Tyr 175 and Glu 177 from another
protein molecule. Besides, the phenyl group of Phe 294 hangs
over the AMC fluorophore in a T-shaped stacking and the
packed fluorophores are sandwiched by the arginine and Phe
294 at each side.
As shown in the enlarged view of the binding site in
monomer (Figure 4D), the substrate is trapped by a loop
(SIRT3−291−299) with its ALY protruding into the active site.
The loop between β6 and α10 on the small domain was defined
as loop A (Supporting Information Figure S4). The residues on
loop A formed four H-bonds with the peptide (Val 292, Gly
295, Glu 296, and Leu 298). These H-bonds are strong drive
forces that entrap the substrate to the protein cavity. His 248 is
conserved in mammal sirtuins and critical for the deacetylase
activity.37−39 The acetyl group of the substrate lysine is in
between Phe 180 and His 248. Meanwhile, residues of Ile 291
and Ile 230 (both are conserved in mammal sirtuins) surround
the methyl group of the ALY through van der Waals
767
dx.doi.org/10.1021/jm301032j | J. Med. Chem. 2013, 56, 761−780

Journal of Medicinal Chemistry
Article
Figure 5. Structure comparison of SIRT3 protein in complex with AceCS2-Kac peptide (RSGKacVMR) and ac-RHKKac-AMC peptide. (A) The
structure alignment of SIRT3 in complex with two peptides. The “SIRT3/ac-RHKKac-AMC” structure is colored in green with the peptide in stick
representation and carbon atoms colored in purple. The “SIRT3/AceCS2-Kac” structure is colored in slate with the carbon atoms of the peptide
colored in orange. In the loop region (SIRT3−159−167), no density map was observed in the “SIRT3/ac-RHKKac-AMC” structure. Loop A and
loop B on the ALY binding site were indicated with arrows (red arrows for “SIRT3/AceCS2-Kac” and orange arrows for “SIRT3/ac-RHKKac-AMC”).
(B) An enlarged view of the binding site in the structure of “SIRT3/ac-RHKKac-AMC”. The peptide forms four H-bonds with the residues on the
small domain. (C) An enlarged view of the binding site in the structure of “SIRT3/AceCS2-Kac”. The peptide forms four H-bonds with the residues
on loop A and three H-bonds with that on loop B. The distances between the acetyl-lysine and loop B of the “SIRT3/AceCS2-Kac” structure were
shorter than that of the “SIRT3/ac-RHKKac-AMC” structure. (D) A lateral view of the binding site in (B). The protein molecule is in electrostatic
surface representation. (E) A lateral view of (D).
interactions. In the structure, the binding cavity is filled with
water molecules as shown in Figure 4D. The water not only
formed H-bond interactions between each other but also
interacted with residues Gln 228, His 248, and Glu 325 of
SIRT3, the acetyl group of ALY, and the linker of coumarin.
Structure Comparison of SIRT3 in Complex with
Different Acetyl Peptides. Jin et. al reported the structure
of SIRT3 in complex with its substrate AcsCS2-Kac (PDB ID:
3GLR).40 The sequence of the peptide is TRSGKacVMRRLLR.
Only the region of RSGKacVMR was observed in the crystal
768
dx.doi.org/10.1021/jm301032j | J. Med. Chem. 2013, 56, 761−780

Journal of Medicinal Chemistry
Article
structure. As for other sirtuin−peptide complex, the ALY of the
substrate inserts into the cleft between the small and large
domains. loop A as well as a loop on the large domain (SIRT3−
323−326 in between β7 and α11, defined as loop B) and the
peptides form an enzyme−substrate antiparallel β-sheet. The
AcsCS2-Kac peptide pulls two loops closer to each other in
comparison with those in the apo-SIRT3 structure (PDB ID:
3GLS). This suggests that binding of the acetyl peptide induced
a conformational change of the enzyme.
To examine whether the binding of the fluorophore-labeled
peptide may induce further conformational change of the cleft,
we superimposed the “SIRT3/AMC-acetyl peptide” complex
structure with the “SIRT3/AcsCS2-Kac peptide” complex
structure. The overall structures of SIRT3 in complex with
two peptides are similar. The root-mean-square deviation
(rmsd) of the Cα atoms of SIRT3 in the two complexes was
0.97 Å. As shown in Figure 5A, the main difference was at the
cleft in between loop A and loop B. In the “SIRT3/ac-RHKKac-
AMC” structure, the distance between the two loops increased
and the cleft adopted a more open conformation. We
speculated that the wider formation of the cleft was due to
the steric block of the coumarin. As shown in Figure 5B, the
residues on loop B from the large domain did not form any
direct interaction with the peptide. Only the carbonyl oxygen of
Glu 325 had indirect connection with the peptide via the H-
bond network of the water molecules (Figure 4D). The vertical
orientation of the coumarin seems to repel the hydrogen
bonding between the substrate and the large domain, while for
AcsCS2-Kac peptide in the SIRT3 complex, the residues after
ALY formed three H-bonds with the residues on loop B (Figure
5C). The comparison of the protein surface at the binding sites
clearly showed that the clefts in the complexes adopted
different conformations due to the features of the acetyl
substrates (Figure 5D,E).
Complex Model of SIRT1 with ac-RHKKac-AMC
Peptide. To elucidate the activation mechanism of our
compounds, we built a SIRT1 model using SIRT3 as the
template by homology modeling. The sequence of SIRT1
(244−498) was aligned with SIRT3 (Supporting Information
Figure S4). In the “SIRT3/ac-RHKKac-AMC” complex, no
density map was observed in the region of SIRT3−158−167,
which is a flexible loop on the protein surface. Thus, the
structure at the regions of 121−157 and 168−391 of “SIRT3/
ac-RHKKac-AMC” complex and the loop region (158−167) of
SIRT3-AcsCS2-Kac complex (PDB ID: 3GLR) were used as the
templates. After structural refinement, the coordinates of AMC-
labeled peptide were copied from “SIRT3/ac-RHKKac-AMC”
complex to the SIRT1 model. The “SIRT1/ac-RHKKac-AMC”
complex model was submitted to further structural refinement.
The complex model of SIRT1 was shown in Figure 6A. The
cleft adopted an open conformation as the “SIRT3/ac-RHKKac-
AMC” structure did. As shown in Figure 6B, the peptide is
captured by the loop of SIRT1−411−419 which corresponds to
loop A. Hydrogen bond interactions were formed between the
peptide and the backbone atoms of SIRT1 residues on loop A
(Val 412, Gly 415, Glu 416, and Leu 418). The phenyl ring of
Phe 414 is adjacent to the AMC fluorophore. The acetyl group
of the lysine is sandwiched by the phenyl ring of Phe 297 and
imidazole ring of His 363. The side chains of Ile 347 and Ile
411 form van der Waals interactions with the methyl group of
the ALY. All the residues mentioned above are conserved in
SIRT1 and SIRT3. The binding mode is similar with the
Figure 6. The homology model of SIRT1 in complex with ac-RHKKac-
AMC peptide. (A) The overall structure of the complex model. SIRT1
is represented in cartoon and colored in silver. Loop A and loop B
were indicated with arrows. (B) The binding site of the complex. The
peptide forms four H-bonds with the backbone atoms on loop A.
SIRT3−substrate complex. The substrate did not have direct
interaction with the residues from loop B.
According to the sirtuin catalytic mechanism, the carbonyl
oxygen at the side chain of ALY carries out nucleophilic attacks
to the C1′ position of the N-ribose of NAD+, which initiates the
deacetylation. After that, an O-alkylamidate intermediate was
formed.41 We suppose that the broadening of the cleft and the
relative position shift of residues involved in catalysis may
hinder the hydrolysis step in deacetylation. The fluorophore,
whether it is coumarin or TAMRA, certainly will cause steric
block in the cleft due to its big and rigid body. The increase of
the distance between the two loops and the loss of the H-bonds
between the substrate and the large domain will weaken the
substrate binding. Both the kinetic experiment and ITC
titration demonstrated that the existence of coumarin increased
the Km and Kd values of the peptide to SIRT1. The Km values of
two acetyl substrates (with or without coumarin) were 551.0
and 130.5 μM, respectively (Figure 3), and the Kd values
calculated by titrating the peptide into SIRT1 were 184 and 44
μM, respectively (Supporting Information Figure S1). On the
basis of above analysis, we built the binding modes of the
activators and proposed their activation mechanism.
769
dx.doi.org/10.1021/jm301032j | J. Med. Chem. 2013, 56, 761−780

Journal of Medicinal Chemistry
Article
Figure 7. The binding modes of SIRT1 activators with “SIRT1/ac-RHKKac-AMC”. (A) A superimposed view of three small compounds that bind to
the complex. The fluorophore-labeled peptide is colored in magentas; compounds 20b, 1, and 4 are colored in cyan, yellow, and green, respectively.
(B) An enlarged view of 20b binding mode. The nitrogen atom on the linker forms hydrogen bond with the peptide. The p-hydroxyl on the 3,5-
dibromo-2,4-dihydroxyphenyl fragment H-bonds with the amino group on the side chain of Gln 345, while the o-hydroxyl H-bonds with the linker
between the substrate and the coumarin. (C) An enlarged view of 4 binding mode. (D) An enlarged view of 1 binding mode. The two hydroxyls H-
bond with the sides chains of Arg 446 and His of the substrate, respectively.
SIRT1 Activators Bind in the Broadened Cleft Caused
by the Fluorophore. To explore the binding sites of the
activators, compounds 20b, 1, and 4 were docked into “SIRT1/
ac-RHKKac-AMC” complex model. The poses of 20b, 1, and 4
were superimposed in the complex model as shown in Figure
7A. Three compounds bound in the hydrophobic pocket
formed by the side chains of His, Arg, Lys, and coumarin of the
peptide and the residues at the large domain of SIRT1. All the
compounds apparently have π−π interactions with the
coumarin.
Binding Mode of Diaryl Acylhydrazone Derivatives. In the
“SIRT3/ac-RHKKac-AMC” crystal structure, two coumarins
form π-stacking; therefore, we speculated that the activators
may stack with the coumarin in the similar manner. As shown
in Figure 7B, the 6-bromonaphthalene fragment of 20b
interacts with the coumarin in a “head-to-head” conformation,
while the dibromo-dihydroxybenzaldehyde ring inserts into the
cleft under the fluorophore. We monitored the hydrogen bonds
between 20b and “SIRT1/ac-RHKKac-AMC” complex; the p-
hydroxyl on the dibromo-dihydroxybenzaldehyde ring interacts
with the amino group from the side chain of Gln 345.
According to the SAR, the hydroxyl is crucial in the activation,
and methylation of the group resulted in a complete loss of the
activity (compare 20b and 20c). Thus, the H-bond interaction
is vital for the activation activity of our compounds. In the
mechanism of sirtuin deacetylation, Gln 345 plays an important
role. The carbonyl group on the side chain of Gln 345 may
make an H-bond with 3′-OH of the ribose from NAD+.42,43 We
speculated that the interaction between the compound and Gln
345 as well as other van der Waals interactions may stabilize the
catalytic core and compensate the shifting induced by the
coumarin; as a result, the Vmax was accelerated. For the
explanation of the Km decrease (Figure 3A), we preferred that
the activators caused tighter binding between the fluorophore-
labeled peptide and SIRT1. In the previous publications, 1 and
4 were also proved to decrease the Km values of the
fluorophore-labeled peptides.14,22,25,26 As shown in Figure 7B,
besides the aromatic stacking, the imino group of 20b makes an
H-bond with the substrate, the o-hydroxyl on the dibromo-
dihydroxybenzaldehyde fragment also H-bonds with the
substrate. Through above interactions, compound 20b
increased the binding affinity of the substrate to the enzyme;
770
dx.doi.org/10.1021/jm301032j | J. Med. Chem. 2013, 56, 761−780

Journal of Medicinal Chemistry
Article
a
Table 5. The Kinetic Parameters of the ac-RHKKac-AMC Peptide to SIRT1 Wild-Type and Mutant Proteins
(A)
compound 20b
blank
b
c
d
e
EC1.5 (μM)
Max A (%)
Km (μM)
Vmax/Km
Km (μM)
Vmax/Km
SIRT1
2.7 0.3
12.2 0.6
4.8 0.2
647
237
353
104.8 8.8
184.0 3.7
191.0 12.6
(B)
13.1
1.0
240.4 14.5
262.6 5.1
268.7 22.9
3.1
0.6
1.8
SIRT1Q345A
SIRT1V445A
3.3
compound 1
blank
b
c
d
e
EC1.5 (μM)
Max A (%)
Km (μM)
Vmax/Km
Km (μM)
Vmax/Km
SIRT1
SIRT1R446A
23.6 3.4
87.2 9.1
802
239
111.3 26.7
134.3 14.6
11.4
4.4
240.4 14.5
256.2 10.6
3.1
2.6
a
(A) The experiments were carried out in the absence or presence of 10 μM compound 20b, respectively. (B) The experiments were carried out in
b
the absence or presence of 50 μM compound 1, respectively. EC1.5 value represents the concentration of an activator required to increase the
enzyme activity by 50%. Percentage of maximum activation achieved. Km, the concentration of substrate that results in half-Vmax for an enzymatic
reaction. Vmax, the maximum initial velocity of an enzymatic reaction.
c
d
e
a
Table 6. Structures and Activities of the Imidazo[1,2-b]thiazole Derivatives Identified by the Sirtris Group
a
b
The structures and SIRT1 activity data of the compounds came from table 2 in the ref 23. EC1.5 value represents the concentration of an activator
c
required to increase the enzyme activity by 50%; Percentage of maximum activation achieved.
the existence of 20b in the broadened cleft may stabilize the
complex conformation and facilitate the following catalysis.
We synthesized compounds with varying aromatic fragments
at the P2 part (Tables 1, 2, and 3). The m-biphenyl substitution
(16d) was less active than p-biphenyl (7). The orientation of
the m-biphenyl is not favorable in the π-stacking with the
coumarin. Introduction of electron-withdrawing group such as
trifluoromethyl group (16a) and heterocyclic atoms oxygen
(16c, 16h) or nitrogen (16i, 16j, and 16k) to the biphenyl
fragment maintained the activity as compared with 7. The
diaryl acylhydrazone derivatives with phenyl substitutions at P2
part exhibited weaker activities (9, 16l, 16m, and 16n). In
contrast, naphthalene substitutions significantly improved the
activity. The substitution at the R1 position of the naphthalene
showed great tolerance (20b, 20e, 20f, 20g, 20h, and 20i). The
above SAR at the R1 position is in line with our binding mode.
Naphthalene substitution is preferred, and the substitutions on
the naphthalene ring may improve the π-stacking.
To verify the binding mode, residues Gln 345 and Val 445,
which were shown to have close contacts with 20b, were
771
dx.doi.org/10.1021/jm301032j | J. Med. Chem. 2013, 56, 761−780

Journal of Medicinal Chemistry
Article
mutated for enzymatic activity and kinetic studies. In the
binding mode, Gln 345 exhibited direct interaction with 20b
through the H-bond; in contrast, Val 445 showed much weaker
interaction (van der Waals interaction) with 20b (Figure 7B).
Therefore, Gln 345 was mutated to alanine to eliminate the H-
bond and Val 445 was mutated to alanine to alter the van der
Waals interaction. The activation effects of compound 20b to
mutant proteins were shown in Supporting Information Figure
S5 and listed in Table 5A. For SIRT1Q345A, 20b is about 4-fold
less active to SIRT1Q345A than to wild-type SIRT1 (for SIRT1,
EC1.5 = 2.7 μM; for SIRT1Q345A, EC1.5 = 12.2 μM). 20b could
only activate the enzymatic activity of SIRT1Q345A by ∼2-fold at
the highest dose. For SIRT1V445A, approximately 2-fold change
of EC1.5 was observed (for SIRT1V445A, EC1.5 = 4.8 μM). 20b
could only activate the enzymatic activity of SIRT1V445A by ∼3-
fold. Furthermore, the kinetic parameters (Km and Vmax/Km) of
the fluorophore-peptide were determined in the absence or
presence of 20b, respectively. For SIRT1 wild-type, addition of
20b resulted in a decrease of 135 μM of the Km value (without
20b, Km = 240.4 μM; with 20b, Km = 104.8 μM), while for
SIRT1Q345A and SIRT1V445A, the amount of Km decrease
reduced to 79 and 78 μM, respectively. Besides, after adding 10
μM of 20b, the ratio of Vmax/Km significantly decreased after
mutation of Gln 345 and Val 445 (for SIRT1, Vmax/Km = 13.1;
for SIRT1Q345A, Vmax/Km = 1.0; for SIRT1V445A, Vmax/Km = 3.3).
From the above analysis, we can see that for SIRT1Q345A and
SIRT1V445A mutants, 20b could not lower the Km or increase
compounds are highly dependent on the position of the
heteroatom. For instance, different positions of the oxygen
atom resulted in ∼10-fold change of EC1.5 (32 and 33). This
may be explained by the reason that the position of the
heteroatom could induce different electron distributions of the
π-cloud, which may affect the π−π interaction between the
compound and the fluorophore. Interestingly, changing the 2-
quinoline group to the 8-quinoline group (31) led to a sharp
loss of potency. In our model, the orientation of the 8-quinoline
group is not favorable to form π-stacking with the coumarin.
Binding Mode of 1. Like compound 20b and 4, 1 was
predicted to bind in the same site. As shown in Figure 7D, the
4-hydroxyphenyl ring of 1 and the coumarin are packed in
“head-to-head” conformation; the p-hydroxyl may form H-bond
with the coumarin. The histidine of the substrate is connected
to Arg 446 on the large domain by 3,5-dihydroxyphenyl of 1
through H-bond. To validate the model, the mutagenesis
studies were performed. 1 activated SIRT1 with an EC1.5 of
23.6 μM and ∼8-fold enzymatic activity (Figure 2), which is
consistent with previous studies.44 We mutated Arg 446 to
alanine, and the activation curve against the mutant was also
determined (Supporting Information Figure S6). The max-
imum activation of 1 against SIRT1R446A exhibited a significant
decrease than that of the wild-type, with only ∼2-fold
enhancement. Meanwhile, the EC1.5 value decreased to 87.2
μM. Nevertheless, compound 20b showed no difference in
activating SIRT1 wild-type and SIRT1R446A (Supporting
Information Figure S6A). We assume that the side chain of
Arg 446 may interact with 1 but not 20b. For further validation,
the kinetic parameters of ac-RHKKac-AMC against SIRT1 wild-
type and mutant with or without 1 were determined. As shown
in Table 5B, the Km and Vmax/Km of the substrate for SIRT1
wild-type and SIRT1R446A were similar; the mutation of the side
chain did not change the binding affinity of the substrate. While
in the presence of 50 μM 1, the comparison of the kinetic
parameters in wild-type (Km = 111.3 μM, Vmax/Km = 11.4) and
mutant (Km = 134.3 μM, Vmax/Km = 4.4) indicated a reduced
affinity of SIRT1 to the coumarin-labeled substrate. Notably,
the ratio of Vmax/Km displayed about 3-fold decrease in the
presence of 1 when the Arg 446 was mutated to alanine. The
mutagenesis studies revealed that the side chain of Arg 446
interacted with the hydroxyl of 1, and the binding was
beneficial in the 1 activation.
Vmax/Km as much as that for the wild-type SIRT1. The mutants
reduced the sensitivities to the compound, indicating that the
side chains of Gln 345 and Val 445 take part in the interactions
with 20b.
Binding Mode of Imidazo[1,2-b]thiazole Derivatives. 4 and
5 belong to imidazo[1,2-b]thiazole derivatives. In 2009, the
Sirtris group reported the activities of these derivatives with
various amide groups.23 In Table 6, we listed 13 compounds
from the publication to discuss their SAR in combination with
our model. As shown in Figure 7C, the piperazonomethyl
fragment inserts into the cleft and forms van der Waals
interactions with the loop on the large domain. The
imidazo[2,1-b][1,3]thiazole group lies under the coumarin.
The surfaces of the two aromatic fragments form a T-shaped
stacking. 4 forms a bend at the phenyl group, and the
trimethoxybenzene group forms a sandwich conformation with
coumarin and arginine of the substrate.
On the basis of the above analysis, we may summarize the
common features of the binding modes of the three SIRT1
activators. The aromatic portion of the SIRT1 activator forms
π-stacking with the fluorophore. Heterocyclic or substituted
aromatic compounds are preferred in the stacking, and SIRT1
activators with these fragments showed improved activities.
Such observations were supported by the previous studies in π-
stacking. It was reported that introduction of the heteroatom to
benzene such as nitrogen could create a dipole and reduce the
π-electron cloud in the spatial extent compared with
benzene.45−47 The substituted benzenes form stronger π-
stacking than the unsubstituted ones.48,49 Furthermore, the
other portion of the activators may insert into the cleft under
the fluorophore in a T-shaped stacking or in an extended
conformation, which interact with residues from the large
domain of SIRT1. The two portions of the activators may
stabilize the substrate binding and promote the following
catalysis. It is worth mentioning that the linker between the two
portions is usually composed of 2−4 carbon/nitrogen atoms.
The length of the linker is appropriate to allow the two
In the publication by the Sirtris group, it is said that an
aromatic group might be favorable at the R position because
the cyclopentyl (23) and benzyl (24) exhibited low activities.23
According to the binding mode of 4 (Figure 7C), the
trimethoxyphenyl group forms π−π interaction with the
coumarin in which the aromatic substitutions at R position
are indeed preferred. Substitution of methoxy groups at the
benzene ring improved the activities as compared 4 and 25 to
27. However, the low activity of 26 indicated that 2,4-
dimethoxybenzene substitution is not well tolerated. A
significant level of SIRT1 activation was observed if substituting
the benzene ring (27) with monocyclic pyrimidine substitution
(28). Furthermore, 2-substituted quinoline substitution (29)
improved the potency in comparison with 4, and 3-substituted
quinoline substitution (30) was slightly more active than 29. In
like manner, introduction of heteroatoms such as nitrogen in
the aromatic ring strengthened the stabilization of π-stacking.
Notably, the 2-qunixaline group substitution (5) resulted in a
significant increase of the activity. Besides, the activities of the
772
dx.doi.org/10.1021/jm301032j | J. Med. Chem. 2013, 56, 761−780

Journal of Medicinal Chemistry
Article
portions in the proper conformations to interact with the
substrate or SIRT1 residues.
The Stability, Bioavailability, and Activity of Diaryl
Acylhydrazone Derivatives in Cells. Apart from the above
in vitro studies, we also tested the stability and bioavailability of
our compound in cells. MCF-7 cells were treated with 100 μM
20b and the cultured medium and cells were collected
separately at different time points. Simultaneously, 100 μM
20b was added to fresh medium to serve as control. In Figure 8,
Figure 9. Western blot analysis of compounds 1, 20b, and 20g on
MCF-7 cells.
concentration of 20b decreased from 100 to 20 μM after 6 h
treatment in cells (Figure 8). In vitro, compound 20g showed a
comparable activation activity as 20b (Table 3). Consistently,
the Western blot results indicated that 20g also decreased the
acetyl-p53 on cell level (Figure 9). 20g was slightly more
effective than 20b. Both 20b and 20g showed better activities
than 1.
Figure 8. Stability test of compound 20b. “Control” (squares)
represents the amount of 20b in fresh cell medium. “Sample” (circles)
represents the total amount of 20b in cultured medium and cells. “In
cells” (triangles) represents the amount of 20b in cells.
Implications in the Discovery of Novel SIRT1
Activators. The aim of this study is to understand the
structural basis by which the SIRT1 activators promote the
deacetylation. The study revealed that the acceleration of the
catalysis was dependent on the structural features of the
substrates. For a specific acetyl substrate (X1···Xn)-Kac-Y, if
residue Y comprises an aromatic ring, such as Phe, His, Tyr, or
Trp, the steric bulk of Y would broaden the cleft of the catalysis
domain as the fluorophore did and decrease the binding affinity
of the substrate. In this case, SIRT1 activators may enhance the
binding of substrate and accelerate the deacetylation. Studies by
Dai et al. provided evidence to support the notion.27 4 was
found to activate the deacetylation of ac-RHKKac-F-NH2 and
ac-RHKKac-W-NH2 in a dose−response way but not on the
substrate ac-RHKKac-A-NH2.27 However, in contrast with the
inhibitory effect of 4 on ac-RHKKac-AMC reported by Dai et
al., 4 was observed to activate the deacetylation of AMC-labeled
peptide in our study. Thus, in search of SIRT1 activators for a
specific substrate, the activation mechanism presented here may
be broadly applicative. For instance, among the five acetylated
lysine in human FOXO3,7,51 three of the ε-acetylated lysines
(Lys 259, Lys 290, and Lys 569) are followed by residues
bearing aromatic rings (Tyr 260, Trp 291, and His 570). The
reported SIRT1 activators may accelerate the deacetylation of
native FOXO3. DNA methyltransferase 1(DNMT1) can be
deacetylated by SIRT1 and a Phe is linked to the C-terminus of
acetyl-Lys 1349,52 and the reported compounds might also
have activation effect on DNMT1. The substrate feature-
dependent activation of SIRT1 activators may explain the
controversial results due to the diversity of SIRT1 substrates.
In addition, the structure of “SIRT3/AcsCS2-Kac” complex
implicated that the binding of peptide induced the change of
squares represent the concentrations of 20b in fresh medium at
different time points (Control). Circles represent the total
amount of 20b in cultured medium and cells (Sample).
Triangles represent the amount of 20b in cells (In cells). In
comparison of the two curves (Control and Sample), 20b
remained stable in fresh medium for 120 min (squares), while
in cultured medium the concentration of 20b started to
decrease after 30 min and it decreased to ∼20 μM after 360
min in cell cultures (circles). These results indicated that the
compound is stable in cells for a time period. As for the
bioavailability, 20b in cells achieved the maximum concen-
tration at 30 min (triangles), at which over half of the
compound could get into the cells.
Thus, the functional activity of our compound was further
explored in a cell-based (MCF-7 cells) deacetylation assay.22 In
the experiment, p53 becomes acetylated at Lys 382 after DNA
damage, which was induced by doxorubicin (DOX).50 To
determine the effect of compound 20b on SIRT1 activity,
trichostatin (TSA) was used to inhibit other classes of
deacetylases. A reported SIRT1 activator, 1, was used as
control. As shown in Figure 9, 1 showed no effect on the acetyl-
p53 level at the concentration lower than 100 μM. The acetyl-
p53 level decreased along with the increase of the
concentration of 1. In the sample panel, the acetyl-p53 level
decreased along with the increase amount of 20b. However, no
effect was detected when the concentration of 20b decreased to
60 μM. The high concentrations of our compound used in the
assay may be explained by the stability and bioavailability of the
compound. In the Western blot, the cells were treated with the
compounds for 6 h. The stability assay showed that the
773
dx.doi.org/10.1021/jm301032j | J. Med. Chem. 2013, 56, 761−780

Journal of Medicinal Chemistry
Article
IS). Chemical shifts were reported in parts per million (ppm, δ)
downfield from tetramethylsilane. Proton coupling patterns are
described as singlet (s), doublet (d), triplet (t), quartet (q), multiplet
(m), and broad (br). Low- and high-resolution mass spectra (LRMS
and HRMS) were obtained with electric, electrospray, and matrix-
assisted laser desorption ionization (EI and ESI) produced by
FinniganMAT-95 and LCQ-DECA spectrometers. HPLC analysis
was conducted for all compounds listed in Tables 2−3 on an Agilent
1100 series HPLC with an Agilent Zorbax Eclipse XDB-C18 (4.6 mm
× 50 mm, 5 μm) reversed phase column with two solvent systems
(MeCN/H2O/TFA and MeOH/H2O/TFA). All the assayed com-
pounds displayed a purity of 95−99% in both solvent systems.
General Procedure for Preparation of Esters 13a−k and
21h−i. A vial equipped with a magnetic stir bar was charged with
Pd(PPh3)4 (0.05 mmol), KF·2H2O (3.0 mmol), aryl bromide (1.0
mol), aryl boronic acid (1.2−1.5 mmol), and 4 mL of DME/H2O/
EtOH (7/3/2, v/v/v). Then the vial was capped with septa and heated
at 110 °C for 50 min under microwave irradiation. The reaction
mixture was extracted with EA. The organic layer was washed with
brine, dried over anhydrous Na2SO4, and concentrated under vacuum.
The crude product was purified by flash column chromatography to
afford methyl esters 13a−k and 21h−i.
protein conformation, and the conformational change is
beneficial for the following catalysis. In light of the binding
modes of the above three SIRT1 activators, one can design
compounds which bind directly with the substrate and the
residues around the peptide binding site, such as Phe 414 from
the small domain and Arg 446 from the large domain (Figure
7). The small molecules targeting the binding site could link the
two domains and decrease the Km value of the substrate.
CONCLUSION
■
In this paper, we discovered a new class of small compounds
that activate SIRT1 deacetylation activity by a fluorescence
method. In the process of this project, it was reported that
some putative activators did not directly bind to SIRT1. We
then tried to understand the reason for the phenomenon. First,
we determined the substrate selectivity of the diaryl
acylhydrazone derivatives and 4 by HPLC. Through enzyme
kinetics studies, 20b and 4 were confirmed to decrease the Km
values of the fluorophore-labeled peptide, while none of them
affected the unlabeled peptide. After that, we tried to capture
the structure of the fluorophore-labeled peptide in the enzyme
and determine the changes of enzyme conformation induced by
the fluorophore. To our delight, the cocrystal structure of
SIRT3 in complex with ac-RHKKac-AMC was obtained. In
comparison with the structure of SIRT3 in complex with
fluorophore-unlabeled peptide, the fluorophore was found to
expand the substrate binding cleft and break the hydrogen
bonds between the substrate and the loop on the Rossmann
fold domain. A π-stacking of two coumarins were observed in a
dimeric arrangement of the SIRT3 complexes. Our enzyme
kinetics experiments and ITC data implied that the activators
could be involved in π-stacking with the fluorophore.
Computational analysis was then carried out to model the
SIRT1 structure based on the “SIRT3/ac-RHKKac-AMC”
complex structure and dock the small molecule activators
into the SIRT1 model. The binding modes of 20b, 1, and 4
were consistent with the SAR and mutagenesis results.
Methyl 4-[2-(Trifluoromethyl)phenyl]benzoate 13a. Yield: 55%.
1H NMR (400 MHz, CDCl3) δ 8.10−8.06 (m, 2H), 7.76 (d, J = 8.0
Hz, 1H), 7. (t, J = 7.6 Hz, 1H), 7.51 (t, J = 7.6 Hz, 1H), 7.41 (d, J =
8.0 Hz, 2H), 7.32 (d, J = 7.6 Hz, 1H), 3.95 (s, 3H).
Methyl 4-(5-Cyano-2-hydroxyphenyl)benzoate 13b. Yield: 36%.
1H NMR (300 MHz, CDCl3) δ 8.17 (d, J = 8.4 Hz, 2H), 7.61−7.52
(m, 4H), 7.06 (d, J = 9.0, 1H), 6.06 (br, 1H), 3.96 (s, 3H).
1
Methyl 4-(2H-1,3-Benzodioxol-5-yl)benzoate 13c. Yield: 83%. H
NMR (300 MHz, CDCl3) δ 8.07 (d, J = 8.4 Hz, 2H), 7.58 (d, J = 8.4
Hz, 2H), 7.13−7.08 (m, 2H), 6.90 (d, J = 8.7 Hz, 1H), 6.02 (s, 2H),
3.93 (s, 3H).
1
Methyl 3-Phenylbenzoate 13d. Yield: 77%. H NMR (300 MHz,
CDCl3) δ 8.29 (s, 1H), 8.02 (d, J = 7.8 Hz, 1H), 7.79 (d, J = 7.8 Hz,
1H), 7.63 (d, J = 7.2 Hz, 2H), 7.55−7.43 (m, 3H), 7.38 (t, J = 7.8 Hz,
1H), 3.95 (s, 3H).
1
Methyl 4-(Furan-2-yl)benzoate 13h. Yield: 64%. H NMR (300
MHz, CDCl3) δ 8.05 (d, J = 8.4 Hz, 2H), 7.73 (d, J = 8.4 Hz, 2H),
7.52 (d, J = 1.8 Hz, 1H), 6.79 (d, J = 3.6 Hz, 1H), 6.51 (dd, J = 1.8, 3.6
Hz, 1H), 3.92 (s, 3H).
1
Methyl 4-(Pyridin-4-yl)benzoate 13i. Yield: 79%. H NMR (400
Next, we put forward the activation mechanism of SIRT1
activators on the basis of the complex structures. For a SIRT1
activator, one portion of its structure forms π-stacking with the
fluorophore of the substrate and the other portion may interact
with the large domain of SIRT1. The small molecules link the
substrate and the large domain together to compensate for the
broadened cleft induced by the fluorophore and thus tighten
the binding of the substrate. Moreover, we examined the
activity, stability, and bioavailability of our compound in cells,
which suggests that the compounds may be used as potentially
interesting probes to study the functions of SIRT1. In
summary, we elaborated the activation mechanism of the
activators from the structural point. The complex structures and
the mechanism described here provide the basis for develop-
ment of novel SIRT1 activators and the studies of sirtuin
activator activities in vivo and in vitro.
MHz, CDCl3) δ 8.71 (d, J = 6.0 Hz, 2H), 8.16 (d, J = 8.4 Hz, 2H),
7.71 (d, J = 8.4 Hz, 2H), 7.56 (d, J = 6.0 Hz, 2H), 3.96 (s, 3H).
Methyl 4-(Pyridin-3-yl)benzoate 13j. Yield: 68%. H NMR (400
1
MHz, CDCl3) δ 8.89 (d, J = 2.0 Hz, 1H), 8.65 (dd, J = 1.2, 4.8 Hz,
1H), 8.15 (d, J = 8.4 Hz, 2H), 7.95−7.92 (m, 1H), 7.69−7.63 (d, J =
8.4 Hz, 2H), 7.42 (dd, J = 4.8, 8.0 Hz, 1H), 3.95 (s, 3H).
1
Methyl 4-(Pyridin-2-yl)benzoate 13k. Yield: 42%. H NMR (400
MHz, CDCl3) δ 8.89 (d, J = 2.0 Hz, 1H), 8.65 (dd, J = 1.2, 4.8 Hz,
1H), 8.15 (d, J = 8.4 Hz, 2H), 7.95−7.92 (m, 1H), 7.69−7.63 (d, J =
8.4 Hz, 2H), 7.42 (dd, J = 4.8, 8.0 Hz, 1H), 3.95 (s, 3H).
Methyl 6-(Furan-2-yl)naphthalene-2-carboxylate 21h. Yield:
1
63%. H NMR (300 MHz, DMSO-d6) δ 8.62 (s, 1H), 8.30 (s, 1H),
8.18 (d, J = 8.7 Hz, 1H), 8.08 (d, J = 9.0 Hz, 1H), 8.01−7.96 (m, 2H),
7.87 (d, J = 1.5 Hz, 1H), 7.20 (d, J = 3.6 Hz, 1H), 6.69 (dd, J = 1.5, 3.3
Hz, 1H), 3.92 (s, 3H).
1
Methyl 6-Phenylnaphthalene-2-carboxylate 21i. Yield: 53%. H
NMR (300 MHz, DMSO-d6) δ 8.68 (s, 1H), 8.33 (s, 1H), 8.24 (d, J =
8.7 Hz, 1H), 8.12 (d, J = 8.7 Hz, 1H), 8.04−7.94 (m, 2H), 7.86 (d, J =
7.5 Hz, 2H), 7.54 (t, J = 7.5 Hz, 2H), 7.44 (t, J = 7.5 Hz, 1H), 3.93 (s,
3H).
EXPERIMENTAL SECTION
■
Material. Unless otherwise noted, the reagents (chemicals) were
purchased from commercial sources and used without further
purification. Analytical thin layer chromatography (TLC) was
performed on HSGF 254 (0.15−0.2 mm thickness). Compound
spots were visualized by UV light (254 nm). Column chromatography
was performed on silica gel FCP 200−300. Mircrowave reactions were
carried on CEM reactor. Nuclear magnetic resonance (NMR) spectra
were recorded on a Bruker AMX-300 NMR 300 or 400 MHz (TMS as
6-Bromonaphthalene-2-carbohydrazide 19b. Representive
Procedure for 14a−k, 19a−e, and 22h−i. To a solution of methyl
6-bromonaphthalene-2-carboxylate (513 mg, 1.94 mmol) in 9 mL of
EtOH was added 3 mL of hydrazine hydrate (85%), and the resulting
mixture was stirred under reflux for 3 h. After cooling to room
temperature, the precipitate was filtered, washed with water, and dried
to afford 6-bromonaphthalene-2-carbohydrazide as a white solid in
1
81% yield. H NMR (300 MHz, DMSO-d6) δ 9.95 (s, 1H), 8.44 (s,
774
dx.doi.org/10.1021/jm301032j | J. Med. Chem. 2013, 56, 761−780

Journal of Medicinal Chemistry
Article
1H), 8.28 (s, 1H), 8.03−7.94 (m, 3H), 7.71 (dd, J = 1.8, 8.7 Hz, 1H),
4.58 (s, 2H).
afford 20b as a white solid in 72% yield. HPLC: 96.84%, tR = 10.71. 1H
NMR (400 MHz, DMSO-d6) δ 12.92 (s, 1H), 12.56 (s, 1H), 10.41
(br, 1H), 8.58 (s, 1H), 8.51 (s, 1H), 8.33 (s, 1H), 8.08−8.03 (m, 3H),
7.81−7.74 (m, 2H). EI-MS m/z (M+) 540. EI-HRMS calcd for (M+)
539.8320, found 539.8250.
4-[2-(Trifluoromethyl)phenyl]benzohydrazide 14a. Yield: 75%. 1H
NMR (400 MHz, DMSO-d6) δ 9.87 (s, 1H), 7.89−7.84 (m, 3H), 7.74
(t, J = 7.6 Hz, 1H), 7.64 (t, J = 7.6 Hz, 1H), 7.43−7.38 (m, 3H), 4.53
(s, 2H).
N′-(3,5-Dibromo-2,4-dihydroxybenzylidene)-2′-(trifluoromethyl)-
biphenyl-4-carbohydrazide 16a. Yield: 56%. HPLC: 99.28%, tR
=
4-(5-Cyano-2-hydroxyphenyl)benzohydrazide 14b. Yield: 72%.
1H NMR (400 MHz, DMSO-d6) δ 10.94 (s, 1H), 9.82 (s, 1H), 7.86
(d, J = 8.4 Hz, 2H), 7.76 (d, J = 2.0 Hz, 1H), 7.69−7.63 (m, 3H), 7.10
(d, J = 8.4 Hz, 1H), 4.55 (s, 2H).
10.18. 1H NMR (400 MHz, DMSO-d6) δ 12.91 (s, 1H), 12.45 (s, 1H),
10.41 (s, 1H), 8.49 (s, 1H), 8.00 (d, J = 8.4 Hz, 2H), 7.88 (d, J = 7.6
Hz, 1H), 7.80−7.75 (m, 2H), 7.67 (t, J = 7.8 Hz, 1H), 7.51 (d, J = 8.0
Hz, 2H), 7.46 (d, J = 7.6 Hz, 1H). EI-MS m/z (M+) 556. EI-HRMS
calcd for (M+) 555.9245, found 555.9200.
1
4-(2H-1,3-Benzodioxol-5-yl)benzohydrazide 14c. Yield: 78%. H
NMR (300 MHz, DMSO-d6) δ 9.79 (s, 1H), 7.87 (d, J = 8.4 Hz, 2H),
7.68 (d, J = 8.4 Hz, 2H), 7.32 (d, J = 1.8 Hz, 1H), 7.22 (dd, J = 1.8, 8.1
Hz, 1H), 7.01 (d, J = 8.1 Hz, 1H), 6.07 (s, 2H), 4.54 (br, 2H).
4-(5-Cyano-2-hydroxyphenyl)-N′-(3,5-dibromo-2,4-
dihydroxybenzylidene)benzohydrazide 16b. Yield: 62%. HPLC:
99.04%, tR = 8.85. 1H NMR (300 MHz, DMSO-d6) δ 12.88 (s,
1H), 12.32 (s, 1H), 10.91 (s, 1H), 9.86 (br, 1H), 8.50 (s, 1H), 8.00 (d,
J = 8.1 Hz, 2H), 7.81−7.73 (m, 4H), 7.67 (dd, J = 2.1, 8.4 Hz, 1H),
7.12 (d, J = 8.4 Hz, 1H). MS (ESI, m/z): 528.0 [M − H]−. HRMS
(ESI) calcd for C21H13Br2N3O4Na ([M + Na]+) 551.9170; found
551.9160.
4-(Benzo[d][1,3]dioxol-5-yl)-N′-(3,5-dibromo-2, 4-
dihydroxybenzylidene)benzohydrazide 16c. Yield: 73%. HPLC:
96.09%, tR = 6.89. 1H NMR (300 MHz, DMSO-d6) δ 12.94 (s,
1H), 12.38 (s, 1H), 10.42 (br, 1H), 8.48 (s, 1H), 8.00 (d, J = 8.1 Hz,
2H), 7.85−7.73 (m, 3H), 7.37 (s, 1H), 7.27 (d, J = 8.1 Hz, 1H), 7.04
(dd, J = 1.5, 8.1 Hz, 1H), 6.09 (s, 2H). MS (ESI, m/z): 533.0 [M +
H]+. HRMS (ESI) calcd for C21H13Br2N2O5 ([M − H]−) 530.9191;
found 530.9186.
1
3-Phenylbenzohydrazide 14d. Yield: 85%. H NMR (400 MHz,
DMSO-d6) δ 9.89 (s, 1H), 8.10 (s, 1H), 7.81 (dd, J = 1.6, 7.6 Hz, 2H),
7.74 (d, J = 7.2 Hz, 2H), 7.56−7.48 (m, 3H), 7.40 (t, J = 7.6 Hz, 1H),
4.55 (s, 2H).
1
4-Phenylbenzohydrazide 14e. Yield: 88%. H NMR (400 MHz,
DMSO-d6) δ 9.84 (s, 1H), 7.92 (d, J = 8.8 Hz, 2H), 7.76−7.71 (m,
4H), 7.49 (t, J = 7.6 Hz, 2H), 7.40 (t, J = 7.6 Hz, 1H), 4.52 (s, 2H).
4-(Furan-2-yl)benzohydrazide 14h. Yield: 82%. 1H NMR (400
MHz, DMSO-d6) δ 9.80 (s, 1H), 7.87 (d, J = 8.4 Hz, 2H), 7.81 (d, J =
0.8 Hz, 1H), 7.76 (d, J = 8.4 Hz, 2H), 7.09 (d, J = 3.6 Hz, 1H), 6.63
(dd, J = 1.6, 3.6 Hz, 1H), 4.53 (s, 2H).
4-(Pyridin-4-yl)benzohydrazide 14i. Yield: 74%. 1H NMR (300
MHz, DMSO-d6) δ 9.89 (s, 1H), 8.66 (dd, J = 1.5, 6.0 Hz, 2H), 7.97
(d, J = 8.4 Hz, 2H), 7.89 (d, J = 8.4 Hz, 2H), 7.76 (dd, J = 1.5, 6.0 Hz,
2H), 4.54 (s, 2H).
N′-(3,5-Dibromo-2,4-dihydroxybenzylidene)biphenyl-3-carbohy-
1
drazide 16d. Yield: 64%. HPLC: 95.70%, tR = 23.56. H NMR (400
4-(Pyridin-3-yl)benzohydrazide 14j. Yield: 79%. 1H NMR (400
MHz, DMSO-d6) δ 9.87 (s, 1H), 8.95 (d, J = 2.0 Hz, 1H), 8.60 (dd, J
= 2.0, 4.8 Hz, 1H), 8.17−8.11 (m, 1H), 7.95 (d, J = 8.4 Hz, 2H), 7.83
(d, J = 8.4 Hz, 2H), 7.51 (dd, J = 4.8, 7.2 Hz, 1H), 4.53 (s, 2H).
MHz, DMSO-d6) δ 12.91 (s, 1H), 12.42 (s, 1H), 10.41 (s, 1H), 8.49
(s, 1H), 8.20 (s, 1H), 7.96−7.90 (m, 2H), 7.79−7.74 (m, 3H), 7.65 (t,
J = 7.6 Hz, 1H), 7.53 (t, J = 7.6 Hz, 2H), 7.43 (t, J = 7.2 Hz, 1H). MS
(ESI, m/z): 489.2 [M
+
H]+. HRMS (ESI) calcd for
1
4-(Pyridin-2-yl)benzohydrazide 14k. Yield: 80%. H NMR (400
C20H14Br2N2O3Na ([M + H]+) 510.9269; found 510.9280.
MHz, DMSO-d6) δ 9.87 (s, 1H), 8.69 (d, J = 4.8 Hz, 1H), 8.16 (d, J =
8.4 Hz, 2H), 8.04 (d, J = 8.0 Hz, 1H), 7.98−7.88 (m, 3H), 7.41−7.37
(m, 1H), 4.56 (s, 2H).
N′-(2,4-Dihydroxybenzylidene)biphenyl-4-carbohydrazide 16e.
Yield: 33%. HPLC: 98.13%, tR = 6.43. 1H NMR (400 MHz,
DMSO-d6) δ 11.98 (s, 1H), 11.50 (s, 1H), 9.97 (s, 1H), 8.53 (s,
1H), 8.02 (d, J = 8.4 Hz, 2H), 7.85 (d, J = 8.4 Hz, 2H), 7.76 (d, J = 7.2
Hz, 2H), 7.51 (t, J = 7.6 Hz, 2H), 7.43 (t, J = 7.2 Hz, 1H), 7.32 (d, J =
8.4 Hz, 1H), 6.41−6.31 (m, 2H). EI-MS m/z (M+) 332. EI-HRMS
calcd for (M+) 332.1161; found 332.1163.
N′-(3,5-Dibromo-2-hydroxy-4-methoxybenzylidene)biphenyl-4-
carbohydrazide 16f. Yield: 78%. HPLC: 99.29%, tR = 16.59. 1H NMR
(400 MHz, DMSO-d6) δ 12.98 (s, 1H), 12.55 (s, 1H), 8.55 (s, 1H),
8.06 (d, J = 8.4 Hz, 2H), 7.91 (s, 1H), 7.87 (d, J = 8.4 Hz, 2H), 7.76
(d, J = 7.6 Hz, 2H), 7.51 (t, J = 7.6 Hz, 2H), 7.43 (t, J = 7.6 Hz, 1H),
3.83 (s, 3H). EI-MS m/z (M+) 502. EI-HRMS calcd for (M+)
501.9528; found 501.9501.
N′-(3,5-Dibromo-4-hydroxy-2-methoxybenzylidene)biphenyl-4-
carbohydrazide 16g. Yield: 81%. HPLC: 96.32%, tR = 6.32. 1H NMR
(400 MHz, DMSO-d6) δ 11.99 (s, 1H), 8.61 (s, 1H), 8.08−8.00 (m,
3H), 7.84 (d, J = 8.4 Hz, 2H), 7.75 (d, J = 7.2 Hz, 2H), 7.51 (t, J = 7.2
Hz, 2H), 7.42 (t, J = 7.2 Hz, 1H), 3.83 (s, 3H). EI-MS m/z (M+) 502.
EI-HRMS calcd for (M+) 501.9528; found 501.9524.
N′-(3,5-Dibromo-2,4-dihydroxybenzylidene)-4-(furan-2-yl)-
benzohydrazide 16h. Yield: 55%. HPLC: 97.26%, tR = 14.51. 1H
NMR (300 MHz, DMSO-d6) δ 12.93 (s, 1H), 12.37 (s, 1H), 10.39 (s,
1H), 8.48 (s, 1H), 8.01 (d, J = 8.4 Hz, 2H), 7.90−7.83 (m, 3H), 7.77
(s, 1H), 7.16 (d, J = 3.6 Hz, 1H), 6.67 (dd, J = 1.8, 3.3 Hz, 1H). EI-MS
m/z (M+) 478. EI-HRMS calcd for (M+) 477.9164; found 477.9148.
N′-(3,5-Dibromo-2,4-dihydroxybenzylidene)-4-(pyridin-4-yl)-
benzohydrazide 16i. Yield: 71%. HPLC: 95.71%, tR = 2.27. 1H NMR
(400 MHz, DMSO-d6) δ 12.90 (s, 1H), 12.46 (s, 1H), 10.32 (s, 1H),
8.70 (d, J = 4.4 Hz, 2H), 8.50 (s, 1H), 8.10 (d, J = 8.4 Hz, 2H), 8.01
(d, J = 8.4 Hz, 2H), 7.83−7.77 (m, 3H). MS (ESI, m/z): 488.1 [M −
H]−. HRMS (ESI) calcd for C19H13Br2N3O3Na ([M + Na]+)
511.9221; found 511.9240.
Quinoline-6-carbohydrazide 19a. Yield: 83%. 1H NMR (400
MHz, DMSO-d6) δ 10.01 (s, 1H), 8.97 (dd, J = 1.6, 4.0 Hz, 1H),
8.51−8.43 (m, 2H), 8.15 (dd, J = 1.6, 8.8 Hz, 1H), 8.06 (d, J = 8.8 Hz,
1H), 7.60 (dd, J = 4.0, 8.4 Hz, 1H), 4.68 (br, 2H).
6-Methanesulfonylnaphthalene-2-carbohydrazide 19c. Yield:
84%. 1H NMR (300 MHz, DMSO-d6) δ 10.06 (s, 1H), 8.64 (s,
1H), 8.56 (s, 1H), 8.30−8.26 (m, 2H), 8.07−7.99 (m, 2H), 4.68 (s,
2H).
6-Cyanonaphthalene-2-carbohydrazide 19d. Yield: 93%. 1H
NMR (300 MHz, DMSO-d6) δ 10.06 (s, 1H), 8.64 (s, 1H), 8.52 (s,
1H), 8.21 (d, J = 8.4 Hz, 1H), 8.13 (d, J = 8.7 Hz, 1H), 8.04 (dd, J =
1.5, 8.4 Hz, 1H), 7.86 (dd, J = 1.5, 8.4 Hz, 1H), 4.63 (s, 2H).
6-Methoxynaphthalene-2-carbohydrazide 19e. Yield: 91%. 1H
NMR (300 MHz, DMSO-d6) δ 9.82 (s, 1H), 8.35 (s, 1H), 7.91 (d, J =
9.0 Hz, 1H), 7.86 (s, 2H), 7.37 (d, J = 2.4 Hz, 1H), 7.22 (dd, J = 2.4,
9.0 Hz, 1H), 4.54 (br, 2H), 3.90 (s, 3H).
6-(Furan-2-yl)naphthalene-2-carbohydrazide 22h. Yield: 88%.
1H NMR (300 MHz, DMSO-d6) δ 9.93 (s, 1H), 8.40 (s, 1H), 8.26
(s, 1H), 8.06−8.00 (m, 2H), 7.95−7.89 (m, 2H), 7.85 (d, J = 1.5 Hz,
1H), 7.16 (d, J = 3.3 Hz, 1H), 6.67 (dd, J = 1.8, 3.3 Hz, 1H), 4.57 (s,
2H).
6-Phenylnaphthalene-2-carbohydrazide 22i. Yield: 84%. 1H
NMR (300 MHz, DMSO-d6) δ 9.95 (s, 1H), 8.46 (s, 1H), 8.28 (s,
1H), 8.10 (d, J = 8.7 Hz, 1H), 8.05 (d, J = 10.0, Hz, 1H), 7.96−7.89
(m, 2H), 7.84 (d, J = 7.5 Hz, 2H), 7.53 (t, J = 7.5 Hz, 2H), 7.43 (t, J =
7.5 Hz, 1H), 4.57 (s, 2H).
6-Bromo-N′-(3,5-dibromo-2,4-dihydroxybenzylidene)-2-
naphthohydrazide 20b. Representive Procedure for Com-
pounds 16a−n and 20a−i. 6-Bromonaphthalene-2-carbohydrazide
(100 mg, 0.38 mmol) and 3,5-dibromo-2,4-dihydroxybenzaldehyde
(117 mg, 0.40 mmol) were suspended in 3 mL of EtOH/HAc (20/1,
v/v). The resulting mixture was stirred for 3 h under reflux. The
precipitate was collected by hot filtration and rinsed with EtOH to
N′-(3,5-Dibromo-2,4-dihydroxybenzylidene)-4-(pyridin-3-yl)-
benzohydrazide 16j. Yield: 63%. HPLC: 98.59%, tR = 3.32. 1H NMR
(400 MHz, DMSO-d6) δ 12.91 (s, 1H), 12.43 (s, 1H), 10.41 (s, 1H),
775
dx.doi.org/10.1021/jm301032j | J. Med. Chem. 2013, 56, 761−780

Journal of Medicinal Chemistry
Article
9.00 (d, J = 2.0 Hz 1H), 8.63 (d, J = 4.0 Hz, 1H), 8.50 (s, 1H), 8.23−
8.16 (m, 1H), 8.08 (d, J = 8.4 Hz, 2H), 7.95 (d, J = 8.4 Hz, 2H), 7.78
(s, 1H), 7.54 (dd, J = 4.8, 8.0 Hz, 1H). MS (ESI, m/z): 490.1 [M +
H]+. HRMS (ESI) calcd for C19H14Br2N3O3 ([M + H]+) 489.9402;
found 489.9391.
N′-(3,5-Dibromo-2,4-dihydroxybenzylidene)-6-(furan-2-yl)-2-
1
naphthohydrazide 20h. Yield: 67%. HPLC: 98.20%, tR = 10.79. H
NMR (300 MHz, DMSO-d6) δ 12.97 (s, 1H), 12.53 (s, 1H), 10.41 (s,
1H), 8.55 (d, J = 0.9 Hz, 1H), 8.52 (s, 1H), 8.31 (s, 1H), 8.13 (dd, J =
4.5, 8.7 Hz, 2H), 8.03−7.96 (m, 2H), 7.87 (dd, J = 0.6, 1.8 Hz, 1H),
7.80 (s, 1H), 7.20 (dd, J = 0.6, 3.3 Hz, 1H), 6.69 (dd, J = 1.8, 3.3 Hz,
1H). MS (ESI, m/z): 529.1 [M + H]+. HRMS (ESI) calcd for
C22H14Br2N2O4Na ([M + Na]+) 550.9218; found 550.9246.
N′-(3,5-Dibromo-2,4-dihydroxybenzylidene)-4-(pyridin-2-yl)-
benzohydrazide 16k. Yield: 65%. HPLC: 99.03%, tR = 3.90. 1H NMR
(400 MHz, DMSO-d6) δ 12.93 (s, 1H), 12.44 (s, 1H), 10.38 (s, 1H),
8.72 (d, J = 4.4 Hz, 1H), 8.50 (s, 1H), 8.27 (d, J = 8.0 Hz, 2H), 8.11−
8.04 (m, 3H), 7.94 (t, J = 7.6 Hz, 1H), 7.78 (s, 1H), 7.45−7.40 (m,
1H). MS (ESI, m/z): 490.2 [M + H]+. HRMS (ESI) calcd for
C19H13Br2N3O3Na ([M + Na]+) 511.9221; found 511.9221.
N′-(3,5-Dibromo-2,4-dihydroxybenzylidene)-4-methylbenzohy-
N′-(3,5-Dibromo-2,4-dihydroxybenzylidene)-6-phenyl-2-naph-
1
thohydrazide 20i. Yield: 91%. HPLC: 97.09%, tR = 15.38. H NMR
(300 MHz, DMSO-d6) δ 12.96 (s, 1H), 12.54 (s, 1H), 8.60 (s, 1H),
8.53 (s, 1H), 8.33 (s, 1H), 8.22−8.13 (m, 2H), 8.06−7.95 (m, 2H),
7.89−7.84 (m, 2H), 7.79 (s, 1H), 7.58−7.51 (m, 2H), 7.44 (t, J = 7.2
Hz, 1H). MS (ESI, m/z): 539.1 [M + H]+. HRMS (ESI) calcd for
C24H16Br2N2O3Na ([M + Na]+) 560.9425; found 560.9414.
1
drazide 16l. Yield: 42%. HPLC: 98.90%, tR = 9.18. H NMR (400
MHz, DMSO-d6) δ 12.94 (s, 1H), 12.26 (s, 1H), 10.39 (s, 1H), 8.45
(s, 1H), 7.85 (d, J = 8.0 Hz, 2H), 7.74 (s, 1H), 7.36 (d, J = 8.0 Hz,
2H), 2.39 (s, 3H). EI-MS m/z (M+) 426. EI-HRMS calcd for (M+)
425.9215; found 425.9213.
Protein Expression and Purification. DNA fragment which
encodes SIRT1 (156−664) was inserted into pET28a vector (Nde1/
Xho1) with a hexa histidine-tag at the N-terminal. The plasmid was
transformed into Escherichia coli BL21 (DE3) competent cells. A single
colony was inoculated in the LB medium with 50 μg/mL kanamycin at
37 °C. Then 500 μM IPTG was added to the cell culture when OD600
reached 0.8−1.0 and the protein was induced at 16 °C overnight. The
cells were sonicated with SIRT1 lysis buffer (20 mM Tris, 200 mM
NaCl, 5 mM β-ME, pH 8.0) and loaded to Ni-NTA. After washing
with 40 mM imidazole, the protein was eluted with 500 mM imidazole.
The elution was validated by SDS-PAGE and dialyzed into sirtuin
deacetylation assay buffer in combination with glycerol (25 mM Tris,
137 mM NaCl, 2.7 mM KCl, 1 mM MgCl2, 10% glycerol, pH 8.0).
SIRT1Q345A, SIRT1V445A, and SIRT1R446A plasmid were generated by
site-directed mutagenesis kit (Stratagene, catalogue no 200518). The
mutant was expressed and purified the same as the wild-type SIRT1.
DNA fragment which encodes SIRT3 (118−399) was inserted into
pET21a vector (BamH1/Xho1) with a hexa histidine-tag at the N-
terminal and a TEV cleavage site between the protein sequence and
his-tag. The plasmid was transformed into E. coli BL21 (DE3)
competent cells. A single colony was inoculated in the LB medium
with 100 μg/mL ampicillin at 37 °C. Then 100 μM IPTG was added
to the cell culture when OD600 reached 0.6−0.8 and the protein was
induced at 16 °C overnight. The cells were sonicated with SIRT3 lysis
buffer (25 mM Hepes, 200 mM NaCl, 5 mM β-ME, 5% glycerol, pH
7.5) and loaded to Ni-NTA. After washing with 40 mM imidazole, the
protein was eluted with 250 mM imidazole. The elution was dialyzed
into SIRT3 dialysis buffer (20 mM Tris, 200 mM NaCl, pH8.0) with
TEV protease digestion at 4 °C overnight. The dialyzed protein was
reloaded to Ni-NTA. The flow-through and 5 mM imizadole elution
were collected and loaded to Superdex 200 (GE Healthcare)
equilibrated with SIRT3 dialyze buffer. The peak representing
SIRT3 monomer was collected and concentrated for crystallization.
DNA fragment which encodes SIRT2 (34−356) was inserted into
pGET-4t-1 vector (BamH1/EcoR1) with a GST fusion protein at the
N-terminus. The plasmid was transformed into E. coli BL21 (DE3)
competent cells. A single colony was inoculated in the LB medium
with 100 μg/mL ampicillin at 37 °C. Then 500 μM IPTG was added
to the cell culture when OD600 reached 0.3 and the protein was
induced at 37 °C for 5 h. The protein was purified by GST resin with
SIRT1 lysis buffer and eluted with 10 mM GSH in SIRT1 lysis buffer.
The elution was validated by SDS-PAGE and dialyzed the same as
SIRT1 protein.
N′-(3,5-Dibromo-2,4-dihydroxybenzylidene)-4-(trifluoromethyl)-
1
benzohydrazide 16m. Yield: 32%. HPLC: 98.94%, tR = 13.82. H
NMR (300 MHz, DMSO-d6) δ 12.79 (s, 1H), 12.54 (s, 1H), 10.43 (s,
1H), 8.49 (s, 1H), 8.14 (d, J = 8.1 Hz, 2H), 7.94 (d, J = 8.1 Hz, 2H),
7.80 (s, 1H). EI-MS m/z (M+) 480. EI-HRMS calcd for (M+)
479.8932; found 479.8916.
N′-(3,5-Dibromo-2,4-dihydroxybenzylidene)-3,5-dimethoxyben-
1
zohydrazide 120b. Yield: 54%. HPLC: 98.41%, tR = 9.41. H NMR
(300 MHz, DMSO-d6) δ 12.86 (s, 1H), 12.25 (s, 1H), 10.39 (s, 1H),
8.46 (s, 1H), 7.75 (s, 1H), 7.08 (d, J = 2.4 Hz, 2H), 6.74 (t, J = 2.4 Hz,
1H), 3.82 (s, 6H). MS (ESI, m/z): 471.1 [M − H]−. HRMS (ESI)
calcd for C16H14Br2N2O5Na ([M + Na]+) 494.9167; found 494.9178.
N′-(3,5-Dibromo-2,4-dihydroxybenzylidene)quinoline-6-carbohy-
1
drazide 20a. Yield: 42%. HPLC: 97.19%, tR = 3.59. H NMR (300
MHz, DMSO-d6) δ 12.89 (s, 1H), 12.59 (s, 1H), 10.41 (s, 1H), 9.04
(dd, J = 1.5, 4.2 Hz, 1H), 8.64 (d, J = 1.2 Hz, 1H), 8.60−8.51 (m, 2H),
8.26 (dd, J = 1.5, 8.4 Hz, 1H), 8.16 (d, J = 8.7 Hz, 1H), 7.80 (s, 1H),
7.67 (dd, J = 4.2, 8.1 Hz, 1H). MS (ESI, m/z): 462.0 [M − H]−.
HRMS (ESI) calcd for C17H11Br2N3O3Na ([M + Na]+) 485.9065;
found 485.9083.
6-Bromo-N′-(3,5-dibromo-2-hydroxy-4-methoxybenzylidene)-2-
1
naphthohydrazide 20c. Yield: 79%. HPLC: 97.64%, tR = 8.04. H
NMR (300 MHz, DMSO-d6) δ 12.95 (s, 1H), 12.68 (s, 1H), 8.60 (s,
1H), 8.57 (s, 1H), 8.34 (d, J = 1.8 Hz 1H), 8.11−8.05 (m, 3H), 7.95
(s, 1H), 7.77 (dd, J = 1.8, 8.7 Hz, 1H), 3.84 (s, 3H). EI-MS m/z (M+)
554. EI-HRMS calcd for (M+) 553.8476; found 553.8457.
6-Bromo-N′-(3,5-dibromo-4-hydroxy-2-methoxybenzylidene)-2-
1
naphthohydrazide 20d. Yield: 39%. HPLC: 95.16%, tR = 8.59. H
NMR (300 MHz, DMSO-d6) δ 12.12 (s, 1H), 10.56 (s, 1H), 8.62 (s,
1H), 8.57 (s, 1H), 8.33 (d, J = 1.2H, 1H), 8.09−8.00 (m, 4H), 7.76
(dd, J = 1.5, 8.7 Hz, 1H), 3.83 (s, 3H). EI-MS m/z (M+) 554. EI-
HRMS calcd for (M+) 553.8476; found 553.8484.
N′-(3,5-Dibromo-2,4-dihydroxybenzylidene)-6-(methylsulfonyl)-
1
2-naphthohydrazide 20e. Yield: 52%. HPLC: 96.63%, tR = 5.64. H
NMR (300 MHz, DMSO-d6) δ 12.89 (s, 1H), 12.65 (s, 1H), 10.44 (s,
1H), 8.70 (s, 2H), 8.53 (s, 1H), 8.39−8.33 (m, 2H), 8.15 (dd, J = 0.9,
9.3 Hz, 1H), 8.06 (dd, J = 1.5, 8.7 Hz, 1H), 7.82 (s, 1H), 3.3 (s, 3H).
MS (ESI, m/z): 541.1 [M + H]+. HRMS (ESI) calcd for
C19H13Br2N2O5S ([M − H]−) 538.8912; found 538.8919.
6-Cyano-N′-(3,5-dibromo-2,4-dihydroxybenzylidene)-2-naphtho-
hydrazide 20f. Yield: 37%. HPLC: 95.79%, tR = 10.50. 1H NMR (300
MHz, DMSO-d6) δ 12.88 (s, 1H), 12.63 (s, 1H), 10.46 (s, 1H), 8.68
(d, J = 9.3 Hz, 2H), 8.52 (s, 1H), 8.30 (d, J = 8.4 Hz, 1H), 8.23 (d, J =
8.4 Hz, 1H), 8.13 (d, J = 9.3 Hz, 1H), 7.92 (d, J = 8.7 Hz, 1H), 7.81 (s,
1H). MS (ESI, m/z): 486.1 [M − H]−. HRMS (ESI) calcd for
C19H11Br2N3O3Na ([M + Na]+) 509.9065; found 509.9043.
Screening of Sirtuin Activators and the EC1.5 Determination
of the Compounds. A fluorescent assay was used to screen the
compounds against SIRT1, SIRT2, and SIRT3. The C-terminus of the
acetyl peptide acetyl-Arg-His-Lys-Lys(ε-acetyl) was labeled with 4-
amino-7-methylcoumarin. The ALY was deacetylated by sirtuins.
Then, AMC could be cleaved from the peptide by trypsin. All
compounds were dissolved in DMSO, and the screening concentration
was fixed at 50 μM; compound 1 was used as control. For the
compound screening against SIRT1, the reaction system contained
750 μM NAD+, 31.25 μM ac-RHKKac-AMC, 1 μM SIRT1, and
compounds/DMSO in sirtuin deacetylation assay buffer (25 mM Tris,
137 mM NaCl, 2.7 mM KCl, 1 mM MgCl2, pH8.0). The reactions
N′-(3,5-Dibromo-2,4-dihydroxybenzylidene)-6-methoxy-2-naph-
1
thohydrazide 20g. Yield: 58%. HPLC: 95.11%, tR = 16.07. H NMR
(300 MHz, DMSO-d6) δ 13.00 (s, 1H), 12.45 (s, 1H), 10.39 (s, 1H),
8.50 (s, 2H), 8.00 (d, J = 9.0 Hz, 1H), 7.96 (s, 2H), 7.78 (s, 1H), 7.43
(d, J = 2.4 Hz, 1H), 7.28 (dd, J = 2.7, 9.0 Hz, 1H), 3.92 (s, 3H). MS
(ESI, m/z): 491.2 [M
−
H]−. HRMS (ESI) calcd for
C19H14Br2N2O4Na ([M + Na]+) 514.9218; found 514.9199.
776
dx.doi.org/10.1021/jm301032j | J. Med. Chem. 2013, 56, 761−780

Journal of Medicinal Chemistry
Article
were carried out at 25 °C for 45 min, stopped with 10 mM
nicotinamide, and developed with 0.5 μg/μL trysin for 30 min. The
fluorescence intensity of AMC was measured with a microplate reader
(POLARstar OPTIMA, BMG Labtech); the excitation and emission
wavelengths were set 340 and 490 nm. To examine the selectivity of
the compounds, we tested the activity of the compounds on SIRT2
and SIRT3. The reaction conditions and procedures were the same as
the screening assay on SIRT1 (750 μM of NAD+, 31.25 μM of ac-
RHKKac-AMC, 50 μM of compounds). To achieve the same
enzymatic activity as SIRT1, the concentrations of SIRT2 and
SIRT3 were 4.2 and 3.6 μM, respectively.
The EC1.5 determination was carried out the same as the screening
assay. Compounds were 3-fold diluted in DMSO and added to
reaction system. EC1.5 and maximum activation were calculated by
fitting the data points with the dose−response function in Origin 8.0
(Origin lab).
standard curve, and the deacetylated amount of substrate was
determined. The velocities were calculated by the deacetylated amount
of substrate divided by time. The calculation of Km was the same as
described above.
Crystallization, Data Collection, And Structure Determina-
tion. For “SIRT3/ac-RHKKac-AMC peptide” cocrystallization, the
molar ratio of ac-RHKKac-AMC and SIRT3 protein was 1:1. SIRT3
was concentrated to 7 mg/mL, and the according amount of ac-
RHKKac-AMC peptide powder was added to the protein. The mixture
was incubated at 4 °C overnight and centrifuged for crystallization.
The “SIRT3/ac-RHKKac-AMC” cocrystals were obtained by the
hanging-drop vapor-diffusion method at 15 °C with crystallization
solution containing 0.085 M Hepes-Na (pH 7.5), 8.5% (v/v) isopropyl
alcohol, 17% (w/v) PEG 4000, and 15% (v/v) glycerol anhydrous.
The “SIRT3/ac-RHKKac-AMC” crystals were flash-frozen in liquid
nitrogen. The diffraction data were collected on beamline BL17U at
Shanghai Synchrotron Radiation Facility (SSRF) and processed by
using HKL2000 program.35 The space group of the complex was R32
with one “SIRT3/ac-RHKKac-AMC” complex in the asymmetric unit.
The data set was reprocessed by Scalepack2mtz in the ccp4 suite,36
and the structure was solved by MolRep53 with apo-SIRT3 structure as
the template (PDB ID: 3GLS). The structure refinement was carried
out by Shelx97,54 Refmac5,55 and COOT.56 The detailed information
of diffraction data and refinement statistics were listed in Table 4.
Homology Modeling and Docking Simulation. The sequence
of SIRT1 (244−498) was retrieved from the UniProt database
(UniProtKB code: Q96EB6, http://www.uniprot.org/). SIRT1 model
was built with the crystal structures of “SIRT3/ac-RHKKac-AMC”
(SIRT3−121−157 and SIRT3−168−391) and “SIRT3/AcsCS2-Kac”
(SIRT3−158−167, PDB ID 3GLR) as templates. The sequences of
SIRT1 and SIRT3 were aligned by ClustalW (Supporting Information
Figure S4).57 The aligned sequences were input into Discovery Studio
(Accelrys) and SIRT1 model structure was built by the Build
Homology module. The model of SIRT1 was further validated by
Profiles-3D,58 and the overall compatibility score of the model was
100.37, which was much higher than the expected low score 52.15.
The Ramachandran plot revealed two residues (Leu 404 and Ala 405)
in the disallowed region (Supporting Information Figure S7B), but the
residues are far away from the peptide binding site. Therefore, the
quality of SIRT1 model was acceptable for the following study. The
model of SIRT1 was superimposed into the SIRT3 complex structure,
and the coordinates of ac-RHKKac-AMC were copied from the
complex structure to the SIRT1 model to build SIRT1/ac-RHKKac-
AMC model.
The structures of compounds 20b, 1, and 4 were built with the
Simulate Structure package in Discovery Studio. The structures were
assigned with CHARm forcefield parameters and optimized by energy
minimization. Autodock 4.2 was used to predict the binding modes of
the compounds in SIRT1.59 SIRT1/ac-RHKKac-AMC model structure
was applied with CHARm forcefield. The mol2 files of compounds and
the pdb file of protein were pretreated with AutoDock Tools 1.5.4.60
In the docking procedure, the genetic algorithm (GA) was used. The
number of GA runs was set to 50, population size was set to 200 and
maximum number of top individuals that automatically survive was set
to 2. The grid box was centered at the substrate binding cleft under the
fluorophore. Other parameters were set as default. After running
autogrid and autodock, the possible poses of compounds in SIRT1
were obtained.
Stability Test. The stability of compound 20b was tested on
human breast cancer cells MCF-7. The cells were treated with 20b for
0, 5, 15, 30, 60, 120, and 360 min. The medium was removed at the
end point, and equal volume of acetonitrile was added. Cells were
washed and treated with acetonitrile. The medium and cell lysate were
collected separately for detection. 20b was added to absolute cell
medium and treated the same to serve as control. The samples were
centrifuged and injected onto the Agilent 1260 Infinity HPLC system
(Agilent Technologies, Waldbronn, Germany). Separation was
achieved using a Hypersil GOLD Dim column (100 mm × 4.6 mm
I.D., 3 μm; Thermo Fisher Scientific, Runcorn, Cheshire, UK) at a
flow rate of 1 mL/min. The column was maintained at room
Isothermal Titration Calorimetry. SIRT1 protein was dialyzed
into sirtuin deacetylation assay buffer (25 mM Tris, 137 mM NaCl, 2.7
mM KCl, 1 mM MgCl2, 5% glycerol, pH 8.0). The acetyl peptide was
injected in 20 drops of 2 μL (the first drop was 0.5 μL and removed).
The cell temperature was 26 °C, reference power was 5 μCal/s, initial
delay was 60 s, and the stirring speed was 1000 rpm. All titrations were
performed on a Microcal iTC200 (GE Healthcare). For the Kd
detection of ac-RHKKac-NH2 to SIRT1, 2.5 mM ac-RHKKac-NH2
was titrated to 102 μM SIRT1. For the Kd detection of ac-RHKKac-
AMC to SIRT1, 2.5 mM ac-RHKKac-AMC was titrated to 124 μM
SIRT1. Raw data were integrated and analyzed with Origin 7.0
software (Origin lab). The data was fit to a one-site binding model.
Determination of Km for Acetyl Peptide with or without
AMC. Fluorescent Assay. To determine the Km values of ac-RHKKac-
AMC for SIRT1 wild-type and mutants in the absence and presence of
compounds, the concentrations of NAD+ and enzyme were fixed, with
gradient concentration of peptide. DMSO was used as blank, and
compounds were added as indicated. The procedures were the same as
the above screening assay. The velocities were calculated by the
discrepancies of fluorescence divided by time. The reciprocal data of
velocities were plotted against the reciprocal data of peptide
concentration, and the x intercept equals to negative 1/Km.
HPLC Assay. To determine the Km values of ac-RHKKac-NH2 in the
absence and presence of compounds, SIRT1 was fixed at 1 μM and
NAD+ was fixed at 2 mM, the concentrations of ac-RHKKac-NH2 were
1000, 500, 250, 125, 62.5, 31.25, 15.625, and 7.8125 μM, respectively.
Indicated compounds were added to the reaction system (50 μM 4, 10
μM compound 20b or DMSO for blank). The reactions were carried
out at 25 °C for 30 min and stopped by 10 mM nicotinamide. The Km
values of ac-RHKKac-AMC were determined in the same way as ac-
RHKKac-NH2 except that the concentrations of the peptide were set to
300, 150, 100, and 70 μM, respectively. After the reactions were
stopped, the solution was centrifuged at 13000 rpm for 10 min, and
the supernatant was transferred to glass vials for HPLC detection.
The reactions of ac-RHKKac-NH2 were monitored by ultra
performance liquid chromatography (UPLC). The separation was
performed at 30 °C by using the Acquity UPLC BEH Shield RP 18
column (50 mm ×2.1 mm I.D., 1.7 μm) with the Waters Van Guard
precolumn (5 mm × 2.1 mm I.D., 1.7 μm). The mobile phase
consisted of buffer A (distilled water containing 0.1% TFA) and buffer
B (acetonitrile containing 0.1% TFA) with a flow rate of 0.4 mL/min.
The UV detection wavelength was set at 220 nm. The retention time
of ac-RHKKac-NH2 was 1.7 min (Supporting Information Figure S2A).
The reactions with ac-RHKKac-AMC were monitored by Dionex
UltiMate 3000 HPLC system. The separation was achieved at 30 °C
by using the Zorbax SB-Aq C18 column (250 mm × 4.6 mm I.D., 5
μm, Agilent Corp., USA) with the Alltima C18 guard column (5 mm ×
4.6 mm I.D., 5 μm, Grace Corp., USA). The mobile phase was the
same as above with a flow rate of 1 mL/min. The UV detection
wavelength was set at 325 nm. The retention times of ac-RHKK-AMC
and ac-RHKKac-AMC were 6 and 10.5 min, respectively (Supporting
Information Figure S2B).
To determine the Km values, the peak area of the remaining acetyl
substrate was converted to molar concentration by using the peptide
777
dx.doi.org/10.1021/jm301032j | J. Med. Chem. 2013, 56, 761−780

Journal of Medicinal Chemistry
Article
temperature. A mobile phase of acetonitrile−5 mM ammonium
acetate−formic acid (50:50:0.1, v/v/v) was employed. The effluent
was monitored via ultraviolet detection at 340 nm.
coactivator 1; SAR, structure−activity relationship; AMC, 4-
amino-7-methylcoumarin; ITC, isothermal titration calorim-
etry; ALY, acetyl-lysine; DNMT1, DNA methyltransferase 1;
DOX, doxorubicin; TSA, trichostatin
Western Blotting. Human cancer cell line MCF-7 cells were
seeded at a density of 1 × 105 in 6-well plates with Dulbecco’s
Modified Eagle’s Medium (DMEM) plus 10% fetal calf serum. Cells
were incubated in a humidified atmosphere containing 5% CO2 at 37
°C for 24 h. Gradient concentrations of 1, 20b, or 20g plus 0.1 μM
TSA and 5 μM DOX were added to MCF-7 cells. After incubation for
6 h, cells were collected and lysed by using a kit from Active Motif
(Carlsbad, CA). Total protein concentration was determined by BCA
kit (catalogue no. P0012, Beyotime Institute of Biotechnology,
Shanghai, China). Samples containing 50 μg of protein were loaded
onto a 9% gel. The proteins were transferred to a nitrocellulose
membrane (Whatman, UK) after electrophoresis. Then 5% nonfat
milk in TBST (10 mM Tris pH7.5, 150 mM NaCl, 0.05% Tween-20)
was used to block the nonspecific reactivity. To determine the acetyl-
p53, the membrane was incubated with polyclonal antibody against
acetyl-p53 (Cell Signaling Technology, catalogue no. 2525) at 4 °C
overnight. Immune complexes were detected with horseradish
peroxidase-conjugated goat antirabbit IgG antibody and visualized
with a SuperSignal West Dura Kit (Pierce). To determine the total
p53, the same blot was reprobed and p53 antibody (Cell Signaling
Technology, catalogue no. 2527) was used.
REFERENCES
■
(1) Landry, J.; Sutton, A.; Tafrov, S. T.; Heller, R. C.; Stebbins, J.;
Pillus, L.; Sternglanz, R. The silencing protein SIR2 and its homologs
are NAD-dependent protein deacetylases. Proc. Natl. Acad. Sci. U. S. A.
2000, 97, 5807−5811.
(2) North, B. J.; Verdin, E. Sirtuins: Sir2-related NAD-dependent
protein deacetylases. Genome Biol. 2004, 5, 224−236.
(3) Vaquero, A.; Scher, M.; Lee, D. H.; Erdjument-Bromage, H.;
Tempst, P.; Reinberg, D. Human SirT1 interacts with histone H1 and
promotes formation of facultative heterochromatin. Mol. Cell 2004, 16,
93−105.
(4) Vaquero, A.; Sternglanz, R.; Reinberg, D. NAD+-dependent
deacetylation of H4 lysine 16 by class III HDACs. Oncogene 2007, 26,
5505−5520.
(5) Picard, F.; Kurtev, M.; Chung, N.; Topark-Ngarm, A.; Senawong,
T.; Machado De Oliveira, R.; Leid, M.; McBurney, M. W.; Guarente,
L. Sirt1 promotes fat mobilization in white adipocytes by repressing
PPAR-gamma. Nature 2004, 429, 771−776.
(6) Qiao, L. P.; Shao, J. H. SIRT1 regulates adiponectin gene
expression through Foxo1-C/enhancer-binding protein alpha tran-
scriptional complex. J. Biol. Chem. 2006, 281, 39915−39924.
(7) Brunet, A.; Sweeney, L. B.; Sturgill, J. F.; Chua, K. F.; Greer, P. L.;
Lin, Y.; Tran, H.; Ross, S. E.; Mostoslavsky, R.; Cohen, H. Y.; Hu, L.
S.; Cheng, H. L.; Jedrychowski, M. P.; Gygi, S. P.; Sinclair, D. A.; Alt,
F. W.; Greenberg, M. E. Stress-dependent regulation of FOXO
transcription factors by the SIRT1 deacetylase. Science 2004, 303,
2011−2015.
(8) Bordone, L.; Motta, M. C.; Picard, F.; Robinson, A.; Jhala, U. S.;
Apfeld, J.; McDonagh, T.; Lemieux, M.; McBurney, M.; Szilvasi, A.;
Easlon, E. J.; Lin, S. J.; Guarente, L. Sirt1 regulates insulin secretion by
repressing UCP2 in pancreatic beta cells. PLoS Biol. 2006, 4, 210−220.
(9) Ramsey, K. M.; Mills, K. F.; Satoh, A.; Imai, S. I. Age-associated
loss of Sirt1-mediated enhancement of glucose-stimulated insulin
secretion in beta cell-specific Sirt1-overexpressing (BESTO) mice.
Aging Cell 2008, 7, 78−88.
ASSOCIATED CONTENT
* Supporting Information
Information about the synthesis of the chemical compounds,
molecular modeling, HPLC data, isothermal titration calorim-
etry, enzyme kinetics data, etc. This material is available free of
charge via the Internet at http://pubs.acs.org.
■
S
Accession Codes
The coordinates of “SIRT3/ac-RHKKac-AMC” complex
structure were deposited in protein data bank with the PDB
ID of 4FZ3.
AUTHOR INFORMATION
Corresponding Author
*For D.L.: E-mail: dxl@mail.shcnc.ac.cn. For H.L.: phone, +86-
21-50806600; fax, +86-21-50807088; E-mail, hliu@mail.shcnc.
ac.cn.
■
(10) Rodgers, J. T.; Lerin, C.; Haas, W.; Gygi, S. P.; Spiegelman, B.
M.; Puigserver, P. Nutrient control of glucose homeostasis through a
complex of PGC-1alpha and SIRT1. Nature 2005, 434, 113−118.
(11) Lagouge, M.; Argmann, C.; Gerhart-Hines, Z.; Meziane, H.;
Lerin, C.; Daussin, F.; Messadeq, N.; Milne, J.; Lambert, P.; Elliott, P.;
Geny, B.; Laakso, M.; Puigserver, P.; Auwerx, J. Resveratrol improves
mitochondrial function and protects against metabolic disease by
activating SIRT1 and PGC-1 alpha. Cell 2006, 127, 1109−1122.
(12) Blander, G.; Guarente, L. The sir2 family of protein
deacetylases. Annu. Rev. Biochem. 2004, 73, 417−435.
Present Address
¶Shanghai Institute of Materia Medica, Chinese Academy of
Sciences, 555 Zu-Chong-Zhi Road, Shanghai 201203, China
Author Contributions
▽These authors contributed equally to this work.
Notes
The authors declare no competing financial interest.
(13) Longo, V. D.; Kennedy, B. K. Sirtuins in aging and age-related
disease. Cell 2006, 126, 257−268.
ACKNOWLEDGMENTS
We thank Prof. Tony Kouzarides for kindly providing the
plasmid of SIRT1, Prof. Anita Buttner for kindly providing the
plasmid of SIRT2, and Prof. Clemens Steegborn for kindly
providing the plasmid of SIRT3. The work was supported by
the National Science Fund (grant 21021063, grant 91229204)
and the Science and Technology Commission of Shanghai
Municipality (grant 08JC1422100).
■
(14) Howitz, K. T.; Bitterman, K. J.; Cohen, H. Y.; Lamming, D. W.;
Lavu, S.; Wood, J. G.; Zipkin, R. E.; Chung, P.; Kisielewski, A.; Zhang,
L.-L.; Scherer, B.; Sinclair, D. A. Small molecule activators of sirtuins
extend Saccharomyces cerevisiae lifespan. Nature 2003, 425, 191−196.
(15) Wood, J. G.; Rogina, B.; Lavu, S.; Howitz, K.; Helfand, S. L.;
Tatar, M.; Sinclair, D. Sirtuin activators mimic caloric restriction and
delay ageing in metazoans. Nature 2004, 430, 686−689.
(16) Bauer, J. H.; Goupil, S.; Garber, G. B.; Helfand, S. L. An
accelerated assay for the identification of lifespan-extending
interventions in Drosophila melanogaster. Proc. Natl. Acad. Sci. U. S.
A. 2004, 101, 12980−12985.
̈
ABBREVIATIONS USED
■
KDAC, lysine deacetylase; NAD+, nicotinamide adenine
dinucleotide; PPARγ, peroxisome proliferator activated recep-
tor γ; FOXO, forkhead homeobox type O; BESTO, β cell-
specific SIRT1-overexpressing; Ucp2, uncoupling protein 2;
PGC-1α, peroxisome proliferator activated receptor gamma
(17) Valenzano, D. R.; Terzibasi, E.; Genade, T.; Cattaneo, A.;
Domenici, L.; Cellerino, A. Resveratrol prolongs lifespan and retards
the onset of age-related markers in a short-lived vertebrate. Curr. Biol.
2006, 16, 296−300.
778
dx.doi.org/10.1021/jm301032j | J. Med. Chem. 2013, 56, 761−780

Journal of Medicinal Chemistry
Article
(18) Sun, C.; Zhang, F.; Ge, X. J.; Yan, T. T.; Chen, X. M.; Shi, X. L.;
Zhai, Q. W. SIRT1 improves insulin sensitivity under insulin-resistant
conditions by repressing PTP1B. Cell Metab. 2007, 6, 307−319.
(19) Csiszar, A.; Labinskyy, N.; Pinto, J. T.; Ballabh, P.; Zhang, H.;
Losonczy, G.; Pearson, K.; de Cabo, R.; Pacher, P.; Zhang, C.;
Ungvari, Z. Resveratrol induces mitochondrial biogenesis in
endothelial cells. Am. J. Physiol.: Heart Circ. Physiol. 2009, 297, 13−20.
(20) Nayagam, V. M.; Wang, X.; Tan, Y. C.; Poulsen, A.; Goh, K. C.;
Ng, T.; Wang, H.; Song, H. Y.; Ni, B.; Entzeroth, M.; Stunkel, W.
SIRT1 modulating compounds from high-throughput screening as
anti-inflammatory and insulin-sensitizing agents. J. Biomol. Screening
2006, 11, 959−967.
(21) Mai, A.; Valente, S.; Meade, S.; Carafa, V.; Tardugno, M.;
Nebbioso, A.; Galmozzi, A.; Mitro, N.; De Fabiani, E.; Altucci, L.;
Kazantsev, A. Study of 1,4-dihydropyridine structural scaffold:
discovery of novel sirtuin activators and inhibitors. J. Med. Chem.
2009, 52, 5496−5504.
(22) Milne, J. C.; Lambert, P. D.; Schenk, S.; Carney, D. P.; Smith, J.
J.; Gagne, D. J.; Jin, L.; Boss, O.; Perni, R. B.; Vu, C. B.; Bemis, J. E.;
Xie, R.; Disch, J. S.; Ng, P. Y.; Nunes, J. J.; Lynch, A. V.; Yang, H.;
Galonek, H.; Israelian, K.; Choy, W.; Iffland, A.; Lavu, S.; Medvedik,
O.; Sinclair, D. A.; Olefsky, J. M.; Jirousek, M. R.; Elliott, P. J.;
Westphal, C. H. Small molecule activators of SIRT1 as therapeutics for
the treatment of type 2 diabetes. Nature 2007, 450, 712−716.
(23) Vu, C. B.; Bemis, J. E.; Disch, J. S.; Ng, P. Y.; Nunes, J. J.; Milne,
J. C.; Carney, D. P.; Lynch, A. V.; Smith, J. J.; Lavu, S.; Lambert, P. D.;
Gagne, D. J.; Jirousek, M. R.; Schenk, S.; Olefsky, J. M.; Perni, R. B.
Discovery of imidazo[1,2-b]thiazole derivatives as novel SIRT1
activators. J. Med. Chem. 2009, 52, 1275−1283.
(24) Bemis, J. E.; Vu, C. B.; Xie, R.; Nunes, J. J.; Ng, P. Y.; Disch, J.
S.; Milne, J. C.; Carney, D. P.; Lynch, A. V.; Jin, L.; Smith, J. J.; Lavu,
S.; Iffland, A.; Jirousek, M. R.; Perni, R. B. Discovery of oxazolo[4,5-
b]pyridines and related heterocyclic analogs as novel SIRT1 activators.
Bioorg. Med. Chem. Lett. 2009, 19, 2350−2353.
(25) Kaeberlein, M.; McDonagh, T.; Heltweg, B.; Hixon, J.;
Westman, E. A.; Caldwell, S. D.; Napper, A.; Curtis, R.; DiStefano,
P. S.; Fields, S.; Bedalov, A.; Kennedy, B. K. Substrate-specific
activation of sirtuins by resveratrol. J. Biol. Chem. 2005, 280, 17038−
17045.
(26) Pacholec, M.; Bleasdale, J. E.; Chrunyk, B.; Cunningham, D.;
Flynn, D.; Garofalo, R. S.; Griffith, D.; Griffor, M.; Loulakis, P.; Pabst,
B.; Qiu, X.; Stockman, B.; Thanabal, V.; Varghese, A.; Ward, J.;
Withka, J.; Ahn, K. SRT1720, SRT2183, SRT1460, and resveratrol are
not direct activators of SIRT1. J. Biol. Chem. 2010, 285, 8340−8351.
(27) Dai, H.; Kustigian, L.; Carney, D.; Case, A.; Considine, T.;
Hubbard, B. P.; Perni, R. B.; Riera, T. V.; Szczepankiewicz, B.; Vlasuk,
G. P.; Stein, R. L. SIRT1 activation by small molecules: kinetic and
biophysical evidence for direct interaction of enzyme and activator. J.
Biol. Chem. 2010, 285, 32695−32703.
(28) Bass, T. M.; Weinkove, D.; Houthoofd, K.; Gems, D.; Partridge,
L. Effects of resveratrol on lifespan in Drosophila melanogaster and
Caenorhabditis elegans. Mech. Ageing Dev. 2007, 128, 546−552.
(29) Feige, J. N.; Lagouge, M.; Canto, C.; Strehle, A.; Houten, S. M.;
Milne, J. C.; Lambert, P. D.; Mataki, C.; Elliott, P. J.; Auwerx, J.
Specific SIRT1 activation mimics low energy levels and protects
against diet-induced metabolic disorders by enhancing fat oxidation.
Cell Metab. 2008, 8, 347−358.
(30) Smith, J. J.; Kenney, R. D.; Gagne, D. J.; Frushour, B. P.; Ladd,
W.; Galonek, H. L.; Israelian, K.; Song, J.; Razvadauskaite, G.; Lynch,
A. V.; Carney, D. P.; Johnson, R. J.; Lavu, S.; Iffland, A.; Elliott, P. J.;
Lambert, P. D.; Elliston, K. O.; Jirousek, M. R.; Milne, J. C.; Boss, O.
Small molecule activators of SIRT1 replicate signaling pathways
triggered by calorie restriction in vivo. BMC Syst. Biol. 2009, 3, 31−45.
(31) He, L.; Zhang, L.; Liu, X.; Li, X.; Zheng, M.; Li, H.; Yu, K.;
Chen, K.; Shen, X.; Jiang, H.; Liu, H. Discovering potent inhibitors
against the beta-hydroxyacyl-acyl carrier protein dehydratase (FabZ) of
Helicobacter pylori: structure-based design, synthesis, bioassay, and
crystal structure determination. J. Med. Chem. 2009, 52, 2465−2481.
(32) Boehlow, T. R.; Harburn, J. J.; Spilling, C. D. Approaches to the
synthesis of some tyrosine-derived marine sponge metabolites:
synthesis of verongamine and purealidin N. J. Org. Chem. 2001, 66,
3111−3118.
(33) Napper, A. D.; Hixon, J.; McDonagh, T.; Keavey, K.; Pons, J. F.;
Barker, J.; Yau, W. T.; Amouzegh, P.; Flegg, A.; Hamelin, E.; Thomas,
R. J.; Kates, M.; Jones, S.; Navia, M. A.; Saunders, J. O.; DiStefano, P.
S.; Curtis, R. Discovery of indoles as potent and selective inhibitors of
the deacetylase SIRT1. J. Med. Chem. 2005, 48, 8045−8054.
(34) Sanders, B. D.; Jackson, B.; Marmorstein, R. Structural basis for
sirtuin function: what we know and what we don’t. Biochim. Biophys.
Acta, Proteins Proteomics 2010, 1804, 1604−1616.
(35) Otwinowski, Z.; Minor, W. Processing of X-ray diffraction data
collected in oscillation mode. Macromol. Crystallogr., Part A 1997, 276,
307−326.
(36) Collaborative computational project, number 4. The CCP4
suite: programs for protein crystallography. Acta. Crystallogr., Sect. D:
Biol. Crystallogr. 1994, 50, 760−763.
(37) Sauve, A. A.; Celic, I.; Avalos, J.; Deng, H.; Boeke, J. D.;
Schramm, V. L. Chemistry of gene silencing: the mechanism of NAD+-
dependent deacetylation reactions. Biochemistry 2001, 40, 15456−
15463.
(38) Smith, B. C.; Denu, J. M. Sir2 protein deacetylases: evidence for
chemical intermediates and functions of a conserved histidine.
Biochemistry 2006, 45, 272−282.
(39) Jackson, M. D.; Schmidt, M. T.; Oppenheimer, N. J.; Denu, J.
M. Mechanism of nicotinamide inhibition and transglycosidation by
Sir2 histone/protein deacetylases. J. Biol. Chem. 2003, 278, 50985−
50998.
(40) Jin, L.; Wei, W.; Jiang, Y.; Peng, H.; Cai, J.; Mao, C.; Dai, H.;
Choy, W.; Bemis, J. E.; Jirousek, M. R.; Milne, J. C.; Westphal, C. H.;
Perni, R. B. Crystal structures of human SIRT3 displaying substrate-
induced conformational changes. J. Biol. Chem. 2009, 284, 24394−
24405.
(41) Avalos, J. L.; Bever, K. M.; Wolberger, C. Mechanism of sirtuin
inhibition by nicotinamide: altering the NAD+ cosubstrate specificity
of a Sir2 enzyme. Mol. Cell 2005, 17, 855−868.
(42) Min, J.; Landry, J.; Sternglanz, R.; Xu, R. M. Crystal structure of
a SIR2 homolog−NAD complex. Cell 2001, 105, 269−279.
(43) Avalos, J. L.; Boeke, J. D.; Wolberger, C. Structural basis for the
mechanism and regulation of Sir2 enzymes. Mol. Cell 2004, 13, 639−
648.
(44) Borra, M. T.; Smith, B. C.; Denu, J. M. Mechanism of human
SIRT1 activation by resveratrol. J. Biol. Chem. 2005, 280, 17187−
17195.
(45) Geronimo, I.; Lee, E. C.; Singh, N. J.; Kim, K. S. How different
are electron-rich and electron-deficient pi interactions? J. Chem. Theory
Comput. 2010, 6, 1931−1934.
(46) Mishra, B. K.; Sathyamurthy, N. Pi−pi interaction in pyridine. J.
Phys. Chem. A 2005, 109, 6−8.
(47) Mishra, B. K.; Sathyamurthy, N. Stacking interaction in pyrazine
dimer. J. Theor. Comput. Chem. 2006, 5, 609−619.
(48) Lee, E. C.; Kim, D.; Jurecka, P.; Tarakeshwar, P.; Hobza, P.;
Kim, K. S. Understanding of assembly phenomena by aromatic−
aromatic interactions: benzene dimer and the substituted systems. J.
Phys. Chem. A 2007, 111, 3446−3457.
(49) Sinnokrot, M. O.; Sherrill, C. D. Substituent effects in pi−pi
interactions: sandwich and T-shaped configurations. J. Am. Chem. Soc.
2004, 126, 7690−7697.
(50) Sakaguchi, K.; Herrera, J. E.; Saito, S.; Miki, T.; Bustin, M.;
Vassilev, A.; Anderson, C. W.; Appella, E. DNA damage activates p53
through a phosphorylation−acetylation cascade. Genes Dev. 1998, 12,
2831−2841.
(51) Wang, F.; Chan, C. H.; Chen, K.; Guan, X.; Lin, H. K.; Tong, Q.
Deacetylation of FOXO3 by SIRT1 or SIRT2 leads to Skp2-mediated
FOXO3 ubiquitination and degradation. Oncogene 2012, 31, 1546−
1557.
(52) Peng, L.; Yuan, Z.; Ling, H.; Fukasawa, K.; Robertson, K.;
Olashaw, N.; Koomen, J.; Chen, J.; Lane, W. S.; Seto, E. SIRT1
779
dx.doi.org/10.1021/jm301032j | J. Med. Chem. 2013, 56, 761−780

Journal of Medicinal Chemistry
Article
deacetylates the DNA methyltransferase 1 (DNMT1) protein and
alters its activities. Mol. Cell. Biol. 2011, 31, 4720−4734.
(53) Vagin, A.; Teplyakov, A. MOLREP: an automated program for
molecular replacement. J. Appl. Crystallogr. 1997, 30, 1022−1025.
(54) Sheldrick, G. M. A short history of SHELX. Acta Crystallogr.,
Sect. A: Found. Crystallogr. 2008, 64, 112−122.
(55) Vagin, A. A.; Steiner, R. A.; Lebedev, A. A.; Potterton, L.;
McNicholas, S.; Long, F.; Murshudov, G. N. REFMAC5 dictionary:
organization of prior chemical knowledge and guidelines for its use.
Acta Crystallogr., Sect. D: Biol. Crystallogr. 2004, 60, 2184−2195.
(56) Emsley, P.; Cowtan, K. Coot: model-building tools for
molecular graphics. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2004,
60, 2126−2132.
(57) Thompson, J. D.; Higgins, D. G.; Gibson, T. J. Clustal-W
improving the sensitivity of progressive multiple sequence alignment
through sequence weighting, position-specific gap penalties and weight
matrix choice. Nucleic Acids Res. 1994, 22, 4673−4680.
(58) Bowie, J. U.; Luthy, R.; Eisenberg, D. A method to identify
protein sequences that fold into a known three-dimensional structure.
Science 1991, 253, 164−170.
(59) Morris, G. M.; Goodsell, D. S.; Halliday, R. S.; Huey, R.; Hart,
W. E.; Belew, R. K.; Olson, A. J. Automated docking using a
Lamarckian genetic algorithm and an empirical binding free energy
function. J. Comput. Chem. 1998, 19, 1639−1662.
(60) Morris, G. M.; Huey, R.; Lindstrom, W.; Sanner, M. F.; Belew,
R. K.; Goodsell, D. S.; Olson, A. J. AutoDock4 and AutoDockTools4:
automated docking with selective receptor flexibility. J. Comput. Chem.
2009, 30, 2785−2791.
780
dx.doi.org/10.1021/jm301032j | J. Med. Chem. 2013, 56, 761−780