Biological and biophysical properties of the histone deacetylase inhibitor suberoylanilide hydroxamic acid are affected by the presence of short alkyl groups on the phenyl ring

Article abstract of DOI:10.1021/jm901561u

Inhibition of histone deacetylases (HDACs) leads to growth arrest, differentiation, or apoptosis of tumor cell lines, suggesting HDACs as promising targets for cancer therapy. At present, only one HDAC inhibitor (HDACi) is used in therapy: suberoylanilide hydroxamic acid (SAHA). Here, we describe the synthesis and biological evaluation of a new series of compounds derived from SAHA by substituting short alkyl chains at various positions of the phenyl ring. Such modifications induced variable effects ranging from partial loss of activity to increased potency. Through molecular modeling, we describe a possible interaction between HDAC7 proline 809, a residue that is strictly conserved within class 2 enzymes only, and the amide group of HDACi, while nuclear magnetic resonance experiments indicated that dimethyl m-substitution may stabilize the inhibitor in the active site. Our data provide novel information on the structure - activity relationship of HDACi and suggest new ways for developing second generation SAHA-like molecules.

Full text of DOI:10.1021/jm901561u

J. Med. Chem. 2010, 53, 1937–1950 1937
DOI: 10.1021/jm901561u
Biological and Biophysical Properties of the Histone Deacetylase Inhibitor Suberoylanilide
Hydroxamic Acid Are Affected by the Presence of Short Alkyl Groups on the Phenyl Ring
†
Frederik Oger, Aurelien Lecorgne, Elisa Sala, Vanessa Nardese, Florence Demay, Soizic Chevance,
‡
§
§
†
‡
ꢀ ꢀ
ꢀ
Danielle C. Desravines, Nataliia Aleksandrova, Remy Le Guevel,†,z Simone Lorenzi,^ Andrea R. Beccari,^ Peter Barath,#
ꢀ
ꢀ
§
‡
,†
Darren J. Hart, Arnaud Bondon,r Daniele Carettoni, Gerard Simonneaux, and Gilles Salbert*
ꢀ
†
‡
Equipe SPARTE, UMR CNRS 6026-Universite Rennes 1, Campus Beaulieu, Batiment 13, 35042 Rennes, France, Cedex, Equipe ICMV, UMR
ꢀ
^
CNRS 6226-Universite Rennes 1, Campus Beaulieu, Batiment 10C, 35042 Rennes, France, Cedex, Axxam, San Raffaele Biomedical Science
§
ꢀ
Park via Olgettina 58, 20132 Milan, Italy, Unit of Virus Host-Cell Interactions, UJF-EMBL-CNRS, UMI 3265, 6 rue Jules Horowitz, BP181,
^
38042 Grenoble, France, Cedex 9, ^Dompe SpA Via Campo di Pilre, I-67100 L’Aquila, Italy, Cancer Research Institute, Slovak Academy of
#
ꢀ
Sciences, Vlarska 7, SK-833 91 Bratislava, Slovak Republic, and Equipe RMN-ILP, Universite de Rennes 1, UMR6026 CNRS, Campus de
ꢀ
z
ꢀ
ꢀ
Villejean, Bat 05, 35033 Rennes, France, Cedex. Present address: U522 INSERM, Faculte de Medecine, Campus de Villejean, 35033 Rennes,
France
Received June 2, 2009
Inhibition of histone deacetylases (HDACs) leads to growth arrest, differentiation, or apoptosis of
tumor cell lines, suggesting HDACs as promising targets for cancer therapy. At present, only one
HDAC inhibitor (HDACi) is used in therapy: suberoylanilide hydroxamic acid (SAHA). Here, we
describe the synthesis and biological evaluation of a new series of compounds derived from SAHA by
substituting short alkyl chains at various positions of the phenyl ring. Such modifications induced
variable effects ranging from partial loss of activity to increased potency. Through molecular modeling,
we describe a possible interaction between HDAC7 proline 809, a residue that is strictly conserved
within class 2 enzymes only, and the amide group of HDACi, while nuclear magnetic resonance
experiments indicated that dimethyl m-substitution may stabilize the inhibitor in the active site. Our
data provide novel information on the structure-activity relationship of HDACi and suggest new ways
for developing second generation SAHA-like molecules.
Introduction
cells in culture. Many of these agents are effective in inhibiting
tumor growth in vivo, and some have entered clinical trials as
possible antitumor agents.4 Class 1 and class 2 HDACs are
sensitive to similar inhibitors (HDACi) derived from both
natural sources and chemical synthesis. HDACi can be classi-
fied into six main groups: short chain fatty acids (e.g., butyrate
and valproic acid), hydroxamic acids [e.g., trichostatin A
(TSA), suberoylanilide acid (SAHA), and oxamflatin], epox-
yketones (e.g., trapoxin and HC-toxin), cyclic peptides (e.g.,
apicidin and depsipeptide), epoxides (e.g., depudecin), and
benzamidines (e.g., MS-27-275).5 Class 3 HDACs, however,
are not sensitive to these HDACi, and specific inhibitors such
as sirtinol (Sir2 inhibitor naphtol) have been developed for the
specific inactivation of these particular HDACs.6
HDAC inhibition has recently been clinically validated as a
new therapeutic strategy for cancer treatment with the Food
and Drug Administration approval of suberoylanilide hydro-
xamic acid (SAHA) for the treatment of cutaneous T cell
lymphoma.7 Intense research activities are ongoing in phar-
maceutical and academic laboratories toward improving the
pharmacokinetic and therapeutic indices of current HDACi.
The classic pharmacophores for hydroxamic acid inhibitors
consist of three distinct structural motifs: the zinc-binding
group, a hydrophobic linker, and a cap group.8 The X-ray
structures of a bacterial HDAC homologue, histone deacety-
lase-like protein (HDLP),9 bound to SAHA or TSA, and,
more recently, human HDAC8,10 confirmed that the zinc-
binding group interacts with a Zn2þ ion at the bottom of a
Histone deacetylases (HDACs)a are enzymes that catalyze
lysine deacetylation in various proteinaceous substrates such
as histones. HDACs have been divided into four classes
on the basis of their protein sequence homology with yeast
(Saccharomyces cerevisiae) deacetylase enzymes.1 Briefly,
class 1 HDACs (composed of HDAC1, -2, -3, and -8) are
closely related to the yeast transcriptional regulator Rpd3
(reduced potassium dependency 3). Class 2 HDACs, includ-
ing HDAC4, -5, -7, and -9 (class 2a) and HDAC6 and -10
(class 2b), share domain similarity with yeast histone deace-
tylase I. The class 3 HDACs form the nicotinamide adenine
dinucleotide-dependent Sir2 (silent information regulator or
sirtuin) family. In humans, seven members of the Sir2 family
have been characterized and ordered in four subclasses.2
Recently, a separate class of HDACs (class 4), comprising
only HDAC11 in mammals, was described.1
Because aberrant histone acetylation has been linked to
malignant diseases,3 HDAC inhibitors (HDACi) bear great
potential as new antitumor drugs. Indeed, they can induce
differentiation, growth arrest, and apoptosis of transformed
*To whom correspondence should be addressed. Tel: 33(0)2 23 23 66
25. Fax: 33(0)2 23 23 67 94. E-mail: gilles.salbert@univ-rennes1.fr.
aAbbreviations: HDAC, histone deacetylase; HDACi, HDAC inhi-
bitors; SAHA, suberoylanilide acid; SACs, SAHA cap derivatives; STD,
saturation transfer difference; TSA, trichostatin A; WaterLOGSY,
water-ligand observation with gradient spectroscopy.
r
2010 American Chemical Society
Published on Web 02/09/2010
pubs.acs.org/jmc

1938 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 5
Oger et al.
Figure 1. Structure of SAHA and general structure of the synthesized compounds. Compounds were derived from SAHA by substitution with
alkyl on the phenyl ring and were named SACs for “SAHA alkyl cap derivatives”. SACs conserve the hydroxamic acid function, the length of
the aliphatic chain, and the amide function of SAHA and are only modified on the phenyl ring.
Table 1. List of Synthesized Compoundsa
channel-like active site. The common zinc-binding group of
many HDACi is the hydroxamate moiety, and modifications
of this group have been modestly successful, yielding isosteres
such as benzamide, R-ketoesters, electrophilic ketones, mer-
captoamide, and phosphonates.8,11 The hydrophobic linker
presents the zinc-binding group to the active site, fills the
hydrophobic channel, and positions the cap group at the
entrance of the active site. The cap group could thus establish
interactions with amino acid residues at the rim of the
channel.12 However, in the crystal structure of SAHA in
complex with the HDAC7 catalytic domain (PDB ID:
3C0Z), the cap moiety is not well-defined, suggesting that
addition of substituents in this part may improve binding and
thus efficiency of the HDACi. Hence, targeting of the cap
group represents an alternative approach to discovering more
potent and/or more selective HDACi such as suggested by
Wiest and co-workers using molecular dynamics simulation
assays.13
Here, we describe the synthesis of a new series of SAHA
derivatives harboring alkyl substituents at different positions
of the phenyl ring. Our data demonstrate that the biological
activity and the biochemical properties of SAHA can be
finely modulated by the differential positioning of these
substituents.
Results
Chemistry. SAHA cap derivatives (SACs, Figure 1)
synthesized and used in this study are depicted in Table 1.
All SACs and SAHA were synthesized using the method
described by Mai et al.14 The synthetic route to SAHA
analogues is shown in Figure 2. First, suberic anhydride 14
was prepared from a mixture of suberic acid and acetic
anhydride in good yield (98%). Then, suberic anhydride
was condensed with various aniline derivatives to introduce
diversity. In all of the syntheses, we noticed the formation of
the expected amide together with the dianilide byproduct
(about 10-40%). The latter was removed from solution by
precipitation and filtration. The suberanilic acid (1a-13a)
was subsequently treated with hydroxylamine in methanol to
give the expected SAHA derivatives (1-13). Depending on
the nature of the aromatic ring, the compounds were ob-
tained in 14-95% yields. All of the compounds were char-
1
acterized by H NMR, 13C NMR, and mass spectrometry.
As expected, 1H NMR spectroscopy mainly showed the
existence of the Z-isomer.15
Biological Evaluation. o-,m-,p-Substitutions of the Phenyl
a The position of alkyl substituents on the SAHA phenyl cap is
indicated.
Ring Differentially Modulate SAHA’s Ability To Inhibit
HDACs in Caco-2 Cells. As HDACi are promising drugs
to treat colorectal cancer16,17 (www.clinicaltrials.gov),
we analyzed SACs activity in Caco-2 colon cancer cells
(Figure 3A,B). Cells were treated with SACs, and the total
HDAC activity was then measured. As shown in Figure 3A,
the position of substitution impacted the effectiveness
of compounds. Indeed, although p-substituted compounds

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 5 1939
Figure 2. Reagents: (i) Acetic anhydride, 140 ꢀC, 1 h.31 (ii) THF, room temperature, 30 min. (iii) Ethyl chloroformiate, triethylamine
(THF, 0 ꢀC, and 10 min) and then hydroxylamine hydrochloride and potassium hydroxide (MeOH, 0 ꢀC, and 15 min).
(4 and 7) were 2-fold less active than SAHA, m-substituted
compounds (3 and 6) were approximately 1.5-fold more
active than SAHA. On the contrary, o-substituted com-
pounds (2 and 5) were 5-10 times less active than SAHA.
We next produced disubstituted compounds. Given that
monomethyl and monoethyl substitution of SAHA showed
similar effects, only dimethylated SACs were subsequently
synthesized. As shown in Figure 3B, the o-substituted com-
pound (11) was totally inactive toward HDAC inhibition,
confirming that o-substitution strongly affects SAHA effi-
ciency. HDAC inhibition was partially restored when one
o-substitution was combined with one m- or p-substitution
(compounds 8-10). Interestingly, the dimethyl-substituted
compounds without o-substitution (12 and 13) were 1.2-
3-fold more active than SAHA, respectively. Taken together,
these results indicate that alkyl substitution of the SAHA
phenyl ring has a profound influence on the effectiveness of
the inhibitors.
Position of Alkyl Substituents in SACs Both Quantitatively
and Qualitatively Influences Protein Acetylation Levels in
Caco-2 Cells. Next, indirect enzyme-linked immunosorbent
assays (ELISAs) were performed to monitor the effects of
SACs on total protein acetylation levels in Caco-2 cells
(Figure 4).The ability of SACs to regulate protein acetylation
levels in these cells was in good agreement with their ability to
inhibit HDACs (compare Figures 3 and 4). Because different
proteins could account for the acetylation signal detected in
this assay, the protein acetylation profiles of Caco-2 cells
treated with SACs were analyzed by Western blots using
antiacetyl lysine antibodies. As shown in Figure 5A, SAHA
induced hyperacetylation of proteins previously identified by
Sehested and co-workers as R-tubulin and histones.18 As
expected, previously identified active compounds (1, 3, 4, 6,
and 7) induced acetylation of these proteins to the same
extent as SAHA. Conversely, no o-substituted compounds,
except 9 and to a lower extent 10, were able to induce protein
hyperacetylation (Figure 5A,B). Compounds 12 and 13 both
induced R-tubulin and histone acetylation, although 13
Figure 3. Ex vivo HDAC inhibition in Caco-2 cells in response to
treatment with (A) monomethyl- and monoethyl-substituted SACs or
(B) dimethyl-substituted SACs. Caco-2 cells were treated in triplicate
using increasing concentrations from 0.1 to 50 μM (six con-
induced only a weak R-tubulin acetylation as compared to
centrations) of the indicated compounds during 4 h. Data are shown
SAHA (Figure 5B), suggesting that this particular com-
as a ratio of IC50 values (IC50SAHA/IC50compound). Results are the
pound could have a different selectivity than SAHA. These
results were directly confirmed by Western blot analysis
mean of at least three independent experiments except for compound
2 (twoindependentexperiments). Error barsare standard errorsof the
mean (SEM) except for compound 2 (standard deviation, SD).
using antiacetylated R-tubulin and antiacetylated histone
H3/H4 antibodies (Supporting Information, S1, S2). Similar
results were obtained in the hepatoma cell line HepG2
(Supporting Information, S3).
Profiling of SACs Efficiency on HDAC1, HDAC3,
HDAC5, HDAC6, and HDAC7. To determine the potential

1940 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 5
Oger et al.
substituted on the SAHA phenyl ring.10,19-21 Remarkably,
the dimethyl m- and p-substituted compound (12) and
particulary the dimethyl m-substituted compound (13) were
more potent on class 1 HDACs and on HDAC7, whereas
their o-dimethyl counterpart (11) was the least active com-
pound of the series on all tested HDACs. Because of the very
high differential activity between 11 and 13, we decided to
explore further these compounds in structure-activity rela-
tionship (SAR) studies.
Molecular Modeling Reveals Proline 809 as a Potential
Pharmacophore for the Design of Class 2 HDAC Selective
Inhibitors. From the series of recombinant HDACs that we
used for in vitro assay of SACs activity, HDAC7 is the only
enzyme for which crystallographic data are available. Hence,
to address structural features related to the contrasting
biological efficiency of o- and m-dimethyl-substituted
compounds (11 and 13, respectively), molecular modeling
analyses were performed using SAHA, 11, and 13 with
HDAC7 (PDB ID: 3C0Z) as a template. As depicted in
Figure 6A, SAHA was efficiently docked in the HDAC7
active site in a position allowing both interaction of the
hydroxamic acid group with a Zn2þ atom and engagement of
the NH from the amide group in a hydrogen bond with the
carbonyl group of P809. Such a positioning of the HDACi is
likely to be favored by interaction of the phenyl ring with
F737. Interestingly, 13 was predicted to establish a similar
hydrogen bond with P809, whereas 11, most likely due to
conformational constraints, had its NH group tilted away
from P809 (Figure 6A). Through the analysis of the mole-
cular dynamics runs of the three complexes, it appears that
the lack of interaction between the NH group of 11 and the
carbonyl group of the P809 may affect the interactions of the
entire molecule with the protein active site. In particular, the
hydroxamic moiety of 11 was predicted to adopt a different
conformation in the zinc-binding pocket as compared to that
of SAHA and 13 (Figure 6B). The root mean squared
deviation (rmsd) of all three compounds calculated on the
heavy atoms of truncated SAHA suggests that the binding of
11 is less stable than the other two ligands (Figure 6A).
SAHA and 13 have a lower rmsd (calculated on the common
structure atoms) than 11 and place their hydroxamic acid
moiety in a very similar way as compared with TSA (not
shown).
Similarly, molecular modeling of SAHA, 11, and 13 in the
active site of the other class 2a enzyme HDAC4 (PDB ID:
2VQW) suggested hydrogen bonding between the amide
NH of SAHA and 13 and the carbonyl group of
P942 (Supporting Information, S7). Accordingly, P942 of
HDAC4 as well as P809 of HDAC7 are likely to be strong
determinants of ligand binding. Examination of the solved
HDAC4 and HDAC7 crystal structures in complex with
HDACi revealed that these proline residues are indeed
involved in the binding of compounds, although through
bridging water molecules (Supporting Information, S8).
Interestingly, proline at this position is only present in class
2 HDACs and thus provides a possible class-specific phar-
macophore for drug targeting.
Figure 4. Quantitative analysis of protein acetylation levels in
Caco-2 cells in response (A) to monomethyl- and monoethyl-
substituted SACs treatments and (B) to dimethyl-substituted SACs
treatments. Caco-2 cells were treated with a 5 μM concentration of
the indicated compounds during 6 h, and indirect ELISA was
performed in triplicate using 1 μg of whole cell protein extracts
and anti acetyl-lysine as the primary antibody. Results are expressed
as fold acetylation corresponding to protein acetylation level mea-
sured in response to indicated compounds as compared to the
protein acetylation level measured using vehicle (DMSO). Error
bars are standard deviations (SDs).
selectivity or specificity of SACs toward individual HDACs,
we performed in vitro experiments with various recombinant
HDACs. Two class 1 (HDAC1 and -3), two class 2a
(HDAC5 and -7), and one class 2b (HDAC6) HDACs were
used for these assays (dose-response curves are detailed in
the Supporting Information, S4-S6). As shown in Table 2,
all o-substituted compounds were less effective than SAHA
at inhibiting class 1, 2a, and 2b HDACs. These data indicate
that a single o-substitution is necessary and sufficient to
dramatically reduce SAHA efficiency toward class 1 and
class 2 HDACs. Conversely, methyl m- and p-substituted
compounds (3 and 4) and ethyl m-substituted compound (6)
were 1.3-3-fold more potent than SAHA (Table 2). Ethyl
p-substituted compound (7) was as active as SAHA on all
HDACs, except on HDAC7 (four times more efficient),
suggesting some selectivity of this compound. Taken
together, these data indicate that short alkyl chains, m- or
p-substituted, may selectively potentiate SAHA-mediated
inhibition of HDAC activity in vitro. Interestingly, crystal-
lographic structures have revealed that the HDAC active site
is surrounded by hydrophobic residues, which could provide
targets for hydrophobic interactions involving alkyl chains
Substitution in m-Position Increases Compound Affinity for
Recombinant HDAC7. Activity assays on recombinant
HDAC enzymes with SACs indicated that o-substituted
compounds (e.g., 11) were 5-100-fold less efficient than
SAHA, whereas dimethyl m- and/or p-substituted com-
pounds (e.g., 13) harbor not only a higher efficiency when
compared to SAHA but also a partial enzymatic selectivity.

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 5 1941
Figure 5. Protein acetylation profile in the soluble fraction of Caco-2 cells in response to (A) monomethyl- and monoethyl-substituted SACs
treatments and (B) dimethyl-substituted SACs treatments. Caco-2 cells were treated with a 5 μM concentration of the indicated compounds
during 6 h. R-Tubulin was used as a loading control. This result is representative of two independent experiments. When cells were treated with
DMSO (1%) as a negative control, only two bands lower than 20 kDa were detectable (bands 2 and 3). The intensity of these bands was
arbitrarily fixed to 1 to perform a densitometric analysis in comparison to SACs treatment. SAHA increased acetylation of at least one protein
close to 50 kDa. As this signal was not detectable in the negative control, the intensity of this band was arbitrarily fixed to 1.
Although molecular modeling suggested structural clues for
the explanation of the lack of activity of 11, these analyses
did not differentiate 13 from SAHA in terms of interaction.
Thus, in an attempt to assess the SAR of compounds 11 and
13 in vitro, we performed saturation transfer difference
(STD)22 and water-ligand observation with gradient
spectroscopy (WaterLOGSY)23 NMR experiments using
purified recombinant HDAC7 (aa 483-903). STD and
WaterLOGSY sequences are known to be both particularly
efficient for the screening of new molecules24,25 and can be
successfully applied to the analysis of mixtures of different
ligands. Both methods rely on the existence of a strong
negative nuclear Overhauser effect (NOE) associated with
large molecules, in contrast with the weak positive NOE
observed for free small molecules in solution. Hence, each
ligand bound to a protein behaves as its large biomolecule
partner and suffers long correlation time and negative NOE.
STD corresponds to a direct transfer of magnetization

1942 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 5
Oger et al.
Table 2. SACs IC50 Values on Representative Class 1 (HDAC1 and HDAC3), Class 2a (HDAC5 and HDAC7), and Class 2b (HDAC6) HDACsa
IC50 ( SEM (μM)
class 1
class 2a
class 2b
compounds
HDAC1
HDAC3
HDAC5
HDAC7
HDAC6
SAHA
1
0.119 ( 0.031
0.071 ( 0.011
0.576 ( 0.286
0.057 ( 0.023
0.076 ( 0.026
1.101 ( 0.149
0.027 ( 0.002
0.149 ( 0.061
1.110 ( 0.240
0.638 ( 0.013
1.142 ( 0.437
11.614 ( 8.836
0.030 ( 0.010
0.017 ( 0.007
0.147 ( 0.053
0.137 ( 0.034
1.348 ( 0.002
0.058 ( 0.032
0.114 ( 0.034
2.263 ( 0.427
0.035 ( 0.005
0.240 ( 0.070
1.750 ( 0.450
1.100 ( 0.202
1.346 ( 0.164
4.989 ( 1.051
0.056 ( 0.004
0.030 ( 0.001
12.60 ( 2.10
16.00 ( 3.30
42.70 ( 4.50
11.30 ( 0.50
18.90 ( 2.70
>100
41.00 ( 10.20
21.60 ( 1.30
>100
0.037 ( 0.003
0.032 ( 0.002
0.251 ( 0.069
0.030 ( 0.010
0.030 ( 0.001
0.215 ( 0.005
0.021 ( 0.001
0.026 ( 0.004
0.206 ( 0.026
0.171 ( 0.002
0.141 ( 0.041
0.615 ( 0.024
0.020 ( 0.001
0.035 ( 0.005
2
3
16.00 ( 1.80
19.00 ( 5.60
>100
4
5
6
15.70 ( 4.10
9.60 ( 0.30
51.20 ( 12.20
37.60 ( 6.10
54.30 ( 18.30
>100
16.10 ( 2.60
10.50 ( 1.60
>100
7
8
9
45.40 ( 11.60
>100
>100
10
11
12
13
12.40 ( 0.50
14.30 ( 7.60
15.90 ( 1.00
18.30 ( 7.30
a IC50 values were determined from a minimum of two independent experiments. In each experiment, dose-response curves of compounds were tested
in quadruplicate. The average IC50 values and the SEMs are reported in micromolar units.
Figure 6. Molecular modeling of interaction between SAHA, compounds 11 and 13, and the HDAC7 catalytic site. (A) SAHA (yellow) and
compounds 11 (magenta) and 13 (green) are depicted in the HDAC7 active site (gray). Amino acids of the HDAC7 active site are indicated. The
putative H-bond interaction between the proton of NH belonging to the amide function and the carbonyl group of proline 809 (P809) of
HDAC7 is depicted as a green dotted line. (B) Close-up view of hydroxamic acid function. The table indicates the rmsd for each tested
compound. The minimized average structure of SAHA (yellow) in complex with HDAC7 protein (gray) has been reported and is shown with
the superimposition of 11 (magenta) and 13 (green) and belonging to the respective minimized average complex structures.
through the protein protons, whereas WaterLOGSY uses
magnetization transfer through water molecules present at
the protein-ligand interface. The results were visualized as a
difference of spectra with and without magnetization trans-
fer and are displayed in Figure 7 for SAHA, 11, and 13.
Interestingly, the relative intensities of the aromatic reso-
nances were no longer representative of an equimolar mix-
ture in STD and WaterLOGSY assays (Figure 7A, e and
f spectra, respectively). The signals corresponding to the
o- (6.92 ppm) and p- (6.77 ppm) protons of 13 were stronger
than the multiplet at 7.26 ppm related to the four o- and
m-protons of SAHA. The multiplet centered at 7.03 ppm,
which corresponds to three protons (m- and p-) of 11, was
only weakly detectable. Hence, these data indicate that the
differential biological activities of 13, SAHA, and 11 are
likely related to their difference in binding to HDAC active
sites. The weak affinity of 11 for HDAC7 catalytic domain
when compared to 13 was confirmed through analysis of the
methyl protons of both compounds (resonances at 2.13 and
2.01 ppm for 13 and 11, respectively; Figure 7B). Again,
despite being in equimolar concentration with 11 and
SAHA, 13 generated the strongest signal. Hence, on the
basis of all of the NMR data, we can rank affinity of these
three ligands for HDAC7 in vitro as 13 > SAHA . 11.
These results indicate that a combination of two alkyl
substituents in the m-position of the phenyl ring increases the
potency of SAHA, likely through a stabilized interaction
within the active site, and could lead to valuable new

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 5 1943
Figure 7. Expanded regions of one-dimensional spectra of (A) the aromatic protons of SAHA, compounds 11 and 13 (A), and the methyl
protons (B). Experiments were performed in 20 mM phosphate buffer, pH 7.4, 100 mM NaCl, and 10% DMSO-d6 and were recorded at 298 K.
(a) SAHA, (b) compound 11, (c) compound 13, (d) equimolar mixture (60 μM) of the three ligands in the presence of 15 μM HDAC7, (e) STD
spectrum recorded on sample d, and (f) WaterLOGSY spectrum recorded on sample d. All of the signals are presented arbitrary positive for
convenience.
HDACi. In this context, the optimal combination and length
of alkyl chains still remain to be determined. Accordingly,
molecular modeling was performed using 13 docked in
HDAC4 (PDB ID: 2VQW) as a reference and various
compounds harboring from two to six carbon-length
m-substituents (Supporting Information, S9). Results indicated
that the number of different solutions increased with length-
ening of the alkyl chain, revealing unconsistent positioning
of the molecules within the active site. Consistent with the
occurrence of variability in ligand position, the number of
solutions relying on hydrogen bonding between the P942 and
the amide NH decreased with chain length (Supporting
Information, S9). A similar trend was observed when analyz-
ing the position of the hydroxamic moiety of the compounds
in proximity to the zinc ion (Supporting Information, S9).
Hence, short chain derivatives are likely to group in the most
active set of compounds.
Cell Growth Inhibitory Effects of SACs in Caco-2 Cancer
Cells and in Nontumoral Cells. Because differential position-
ing of alkyl substituents on the SAHA phenyl ring differen-
tially impacted HDACi effectiveness in vitro, we performed
cell growth assays (3-[4,5-dimethylthiazol-2-yl]-2,5-diphe-
nyltetrazolium bromide, MTT) to correlate these in vitro
data to the growth inhibition profile of the various com-
pounds in tumor cells. In parallel, we monitored p21waf1/cip1
gene expression, a known SAHA-induced gene regulating
cell cycle and apoptosis.26 Caco-2 cells were treated during
48 h with increasing concentrations of SACs, SAHA, or
vehicle only as a negative control. Mono m-substituted or
p-substituted SACs (3, 4, 6, and 7, respectively) showed IC50
values within the range of SAHA (Figure 8A). As monitored
by Western blot, compounds 3, 4, 6, and 7 also induced
p21waf1/cip1 expression in a dose-dependent manner similarly
to SAHA (Figure 8A). Mono o-substituted SACs (2 and 5)
were less efficient than SAHA with IC50 values approximately
5-fold higher than SAHA (78.5 and 40.5 μM, respectively). In
agreement with these observations, p21waf1/cip1 expression was
barely induced by a 5 μM treatment with these compounds
(Figure 8A). Nonetheless, at a concentration of 50 μM,
o-substituted SACs 2, 5, 8, and 11 were able to induce
p21waf1/cip1 expression (Supporting Information, S10). Taken
together, these results suggest that mono o-substituted SACs
are poor inhibitors of tumoral cell growth.
As shown in Figure 8A, dimethyl-substituted SACs harbor-
ing at least one o-substitution (8-11) exhibited higher IC50
values than SAHA in MTT assay. These high IC50 values were
correlated to a lack of p21waf/cip1 induction. Conversely, m-,
p-dimethyl SACs (12 and 13) were as potent as SAHA in
inhibiting cell growth and induced p21waf1/cip1 to a similar
extent (Figure 8A). Similar results were obtained with HepG2
cells (Supporting Information, S11), albeit 5 and 10 weremore
potent in inducing p21 expression in hepatoma cells than in
Caco-2 cells. These observations indicate a strong correlation
between the ability of these compounds to inhibit HDACs
in vitro and in vivo and their ability to inhibit cell growth in
cancer cells. Although IC50 values from 13 and SAHA were
identical, dose-response curves were different and suggested
that 13 is more potent than SAHA at concentrations ranging
from 12.5 to 50 μM (Figure 8B). Indeed, Hill slopes were
markedly different (-2.03 and -7.16 for SAHA and 13,
respectively), suggesting different modes of interaction with
HDACs. The steepest slope observed with 13 could reflect a
concentration-dependent stabilization of conformational
transitions favoring binding of the inhibitor at high concen-
trations. Alternatively, binding of 13 to HDACs could induce
phase transitions leading to enzyme aggregation, a pheno-
menon that could be interpreted as increased inhibition.
However, slopes observed in in vitro inhibition assays with
recombinant enzymes did not differ signicantly between
SAHA and 13 (see the Supporting Information, S6), suggest-
ing that the steepest slope observed in the presence of 13 in
MTT assay might reflect cell-associated mechanisms.

1944 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 5
Oger et al.
Because 13 was more efficient than SAHA in several in
vitro experiments and was partially more efficient in tumoral
cell growth inhibition, we decided to compare 13 and SAHA
activity toward growth of BJ and MUF nontumoral cells.
Indeed, HDACi are known to be poor inducers of apoptosis
in normal cells.27 As expected, both HDACi had a moderate
effect on nontumoral cell growth (Figure 9). Indeed, SAHA
and 13 IC50 values could not be determined (>100 μM). Cell
growth was similarly affected by SAHA and 13 in BJ and
MUF cells as judged from the almost superimposable do-
se-response curves (Figure 9). These data indicate that 13,
beside being more active than SAHA on several biological
parameters, is not more detrimental than SAHA to normal
cell growth, suggesting that 13 could be a potential candidate
for further investigation.
Discussion
Distinguishing characteristics of HDACi include a metal
binding moiety, a carbon linker, and a capping group.
Considering SAHA, the molecule is characterized as having
an anilide as a capping group, a (CH2)6 chain as a linker, and a
hydroxamic group as a metal-chelating agent. On the basis of
crystallographic analyses, the capping group is solvent-
exposed and could potentially interact with amino acids near
the entrance of the active site.28 Most of the previous HDACi
design have emphasized variations of the capping group
of SAHA derivatives but with large modifications sampling
the surface of the enzyme, such as changing the phenyl
group to a benzyl group and adding functionalities in the
R-position (R-aminosuberic acid).29 In contrast, there is no
systematic study involving discrete modifications of the
phenyl ring while leaving the remainder of the molecule
unchanged.
Our data demonstrate the importance of the position at
which substituents are placed on the phenyl ring of SAHA.
Indeed, o-substituted compounds were mostly inactive when
compared to SAHA, whereas m- and p-substituted com-
pounds were often more active than SAHA. The lack of
potency of o-substituted compounds could be due to their
low affinity for HDAC active sites. Indeed, although HDAC
active sites might harbor a certain structural plasticity, as
suggested by the recent resolution of HDAC4 crystallo-
graphic structure,19 their surface might not be flexible enough
to accommodate o-substitution of the SAHA phenyl ring.10
Hence, the lack of potency of o-substituted compounds might
not be due to steric constraints imposed by the protein moiety.
Accordingly, molecular modeling did not reveal clashes
between 11 and HDAC7 catalytic site. Interestingly, the
collected data rather indicate that o-substitution might favor
a particular compound conformation, altering both Zn2þ
coordination and H-bond formation with the carbonyl group
of P809. This amino acid, which is strictly conserved among
class 2 HDACs only,19 might be essential to stabilize the
binding of HDACi, especially hydroxamic acids such as TSA
and SAHA, in the active site of class 2 enzymes. Indeed,
modeling data for SAHA and 13 indicate that Zn2þ is well-
coordinated by the hydoxamic acid function and that the
amide NH of SAHA and 13 could establish a H-bond with the
carbonyl group of P809. Although molecular modeling al-
lowed us to delineate structural features, which could explain
the weak efficiency of o-substituted compounds, it did not
explain the gain of potency of 13 on various biological
parameters when compared to SAHA. However, STD and
Figure 8. Growth inhibition assays (MTT) and p21waf1/cip1 induction
in response to SACs treatments in Caco-2 cells. (A) Cytotoxicity assays
were performed using SACs concentrations ranging from 0.05 to
100 μM during 48 h of treatment. Results shown are representative
of at least two independent experiments. p21waf1/cip1 induction was
analyzed after 1 or 5 μM SACs treatment during 24 h. IC50
=
compound concentration leading to 50% of viable cells. p21waf1/cip1
expression was monitored using Western blot after 24 h of treatment
with 1 and 5 μM compounds. Equal loading between lanes was verified
using antiactin antibodies (data not shown). (B) MTT dose-response
curves for compound 13 and SAHA-treated Caco-2 cells.

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 5 1945
state and that this effect is amplified as the bulk of the
o-substituent increases. In some cases, such as with 2,4,6-tri-
t-butylacetanilide, a 55/45 (trans/cis) mixture of isomers is
obtained, and the isomers can be separated. It is thus likely
thato-substitutionleads toSACsconformations that preclude
a stable interaction with the enzymes and possibly to a
decrease in the ability of the amide function to establish
critical bonds within its immediate environment.
Conclusion
Our data provide clear evidence that minor modifications
of the phenyl ring of SAHA can significantly change the
properties of the inhibitor. Indeed, we report that the presence
of alkyl groups in o-position (methyl or ethyl) decreases
activity, whereas in m- and p-position, the presence of alkyl
groups leads to a selective increase in activity, as revealed for
12 and 13. In addition, we provide biophysical evidence that
the increase in biological activity of 13 is correlated to an
enhanced affinity of the compound toward the class 2
HDAC7 catalytic domain. Nevertheless, our understanding
of the molecular mechanisms involved awaits structural in-
formation describing interaction between SACs and HDACs
at the atomic level. Our data indicate that, when designing
novel HDACi derived from SAHA, targeting the cap moiety
may provide opportunities to develop more selective and/or
more potent molecules, which could be used in physiopatho-
logical events such as hypercholesterolemia32 or in various
cancers.33
Figure 9. Growth of BJ and MUF nontumoral cells in the presence
of SAHA and compound 13. Cell growth assays were performed
using SAHA and 13 concentrations ranging from 0.05 to 100
micromolar during 48 h of treatment.
Experimental Section
Chemistry. All reactions were performed under argon. Sol-
vents were distilled from the appropriate drying agent prior to
use: tetrahydrofuran (THF) from sodium and benzophenone
and MeOH from magnesium. Commercially available reagents
were used without further purification unless otherwise stated.
All reactions were monitored by TLC with Merck precoated
aluminum foil sheets (Silica gel 60 with fluorescent indicator
UV254). Compounds were visualized with UV light at 254 and
WaterLOGSY NMR experiments indicated that 13 has a
higher affinity for HDAC7 catalytic domain than SAHA.
Indeed, analysis of NMR spectra suggested that the Koff (first-
order dissociation rate constant) of 13 might be lower than the
Koff ofSAHA. Interestingly, although the cap group ofSAHA
is unresolved in the crystal structure of HDAC7, the protons
doshowinteraction withthe proteininSTD experiments. This
interaction seemed to be stronger when methyl groups were
substituted in the m-position (13). The proximity of F737 to
the phenyl ring of 13 could be involved in the strengthening of
this interaction. Cocrystallization of HDAC7 with 13 would
help to define the structural requirements for its binding to the
catalytic site of the enzyme.
1
365 nm. H NMR and 13C NMR were recorded using Bruker
(Advance 300dpx and 200dpx) spectrometers at 300 (or 200
MHz) and 75 MHz (or 50 MHz), respectively. High-resolution
mass spectra were recorded on a ZabSpec TOF Micromass
spectrometer in ESI positive mode (compounds 1-13), on a
Varian MAT 311 Micromass spectrometer in EI mode
ꢀ
(compounds 1a and 3a-13a) at the CRMPO (Centre Regional
de Mesures Physiques de l’Ouest, University of Rennes), and on
a Waters Q-TOF2 Micromass spectrometer in ESI positive
mode (compound 2a). Melting points were determined on a
Kofler melting apparatus. All tested compounds are >95%
pure by elemental analysis (Supporting Information, S12).
All of the syntheses of the SAHA derivatives were adapted
from a method previously reported by Mai et al.14 in three steps.
The preparation of N-hydroxy-N0-(4-methylphenyl)octane-
diamide (AC22) (4) is described as a general example.
Suberic Anhydride (14). A solution of suberic acid (5.0 g, 28.7
mmol) in acetic anhydride (10 mL) was heated at reflux while it
was stirred for 1 h. After it was cooled to room temperature, the
solvent was evaporated under reduced pressure to give 4.40 g
(98%) of a white solid (mp 52 ꢀC). 1H NMR (DMSO-d6), white
solid, mp 52 ꢀC, δ (ppm): 1.34 (m, 4H, H4,5), 1.62 (m, 4H, H3,6),
2.41 (m, 4H, H2,7).
The key structural features of amides, especially short C-N
bond lengths and hindered rotation about the C-N bond,
have been extensively studied.30 It is now well-established that
the E/Z isomerism arises from delocalization of the lone pair
of nitrogen into the π* orbital of the carbonyl group with
some double bond character and slow rotation. The barriers
of rotation of amides have been studied mainly by nuclear
magnetic resonance, but there are few studies of monosub-
stituted amides, due in part to the predominance of the trans
form over the cis form.31 In contrast, data for acetanilide
derivatives are quite abundant, and the amount of cis isomer
increases for acetanilides as the bulk of the ortho-substitution
increases. For example, it has been reported that the trans/cis
isomer ratio is >99 for acetanilide, 94/6 for 2-methylacetani-
lide, and 74/26 for 2,6-dimethylacetanilide. This increase in
the cis isomer is paralleled by an increase in the barrier to
rotation. Although a detailed explanation for the effects of
varying the o-substituent is still missing, it was proposed that
the o-substituted ring is out of the amide plane in the ground
8-[(4-Methylphenyl)amino]-8-oxooctanoic Acid (4a). 4-Methyl-
aniline (343 mg, 3.21 mmol) was added to a stirred solution
of suberic anhydride (500 mg, 3.21 mmol) in anhydrous THF.
After it was stirred at room temperature for 30 min, the result-
ing mixture was filtered to give a solid as the bis-amide

1946 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 5
Oger et al.
(150 mg, 13%). The resulting filtrate was then diluted with water,
yielding a white precipitate that was also filtered to give 370 mg
(44%) of a white solid;mp145ꢀC.1H NMR(DMSO-d6) δ (ppm):
1.29 (m, 4H, 2H4, and 2H5), 1.53 (m, 4H, 2H3, and 2H6), 2.24
(m, 7H, CH3-Ar, 2H2, and 2H7), 7.06 (d, 2H, 2H-Ar), 7.45 (d,
2H, 2H-Ar), 9.78 (s, 1H, NH). 13C NMR (DMSO-d6) δ (ppm):
21.3, 25.3, 25.9, 29.3, 34.5, 37.2, 39.1, 119.9, 129.9, 132.6, 137.7,
171.9, 175.4. MS: Mþ calcd for C15H21NO3, 263.1521; found,
263.1510
Scheme 1. 1H Atom-Labeling Used for SACs
26.0, 26.2, 29.3, 33.2, 36.6, 122.5, 126.7, 127.2, 129.3, 136.7,
139.0, 170.0, 172.3. MS: [M þ Na]þ calcd for C16H24N2O3Na,
315.1685; found, 315.1680
N-Hydroxy-N0-(4-methylphenyl)octanediamide (4). To a 0 ꢀC
cooled solution of 8-[(4-methylphenyl)amino]-8-oxooctanoic
acid (4a) (370 mg, 1.41 mmol) in anhydrous THF, ethyl chloro-
formiate (213 mg, 1.97 mmol) and triethylamine (213 mg,
2.1 mmol) were added, and the mixture was stirred for 10 min.
The solid was filtered off under argon, and the filtrate was added
to freshly prepared solution of hydroxylamine in methanol. To
prepare the hydroxylamine, a solution of hydroxylamine hydro-
chloride (293 mg, 4.22 mmol) in methanol (5 mL) was added to a
stirred solution of potassium hydroxide (236 mg, 4.22 mmol) in
methanol (5 mL) at 0 ꢀC. After it was stirred for 15 min, the
precipitate was removed, and the filtrate was used as such. The
resulting mixture was stirred at room temperature for 15 min
and then was evaporated. After the addition of 10 mL of
acetonitrile to the residue, the solution was filtered to give a
white solid of (4) (390 mg, 44%); mp 152 ꢀC. 1H NMR (DMSO-
d6) δ (ppm): 1.27 (m, 4H, 2H4, and 2H5), 1.52 (m, 4H, 2H3, and
2H6), 1.94 (m, 2H, 2H7), 2.27 (m, 5H, 2H2, and CH3), 7.06 (d,
2H, 2Ar-H), 7.46 (d, 2H, 2Ar-H), 8.69 (s, 1H, OH), 9.80 (s, 1H,
NH), 10.37 (s, 1H, NH). 13C NMR (DMSO-d6) δ (ppm): 21.3,
25.9, 29.3, 33.1, 37.2, 78.6, 119.9, 129.9, 132.6, 137.7, 169.9,
171.8. MS: [M þ Na]þ calcd for C15H22N2O3Na, 301.1528;
found, 301.1526.
N-(3-Ethylphenyl)-N0-hydroxyoctanediamide (6). White solid
(62%), mp 120 ꢀC. 1H NMR (DMSO-d6) δ (ppm): 1.15 (t, 3H,
CH2CH3), 1.26 (m, 4H, 2H4, and 2H5), 1.48 (m, 2H, 1H3, and
1H6), 1.55 (m, 2H, 1H3, and 1H6), 1.93 (m, 2H, 2H7), 2.27 (t, 2H,
2H2), 2.51 (q, 2H, CH2CH3), 6.86 (d, 1H, Ar-H), 7.16 (t, 1H,
Ar-H), 7.40 (d, 1H, Ar-H), 7.45 (s, 1H, Ar-H), 8.65 (s, 1H,
OH), 9.86 (s, 1H, NH), 10.35 (s, 1H, NH). 13C NMR (DMSO-
d6) δ (ppm): 16.5, 26.0, 29.2, 29.3, 33.1, 37.2, 117.4, 119.3, 123.3,
129.4, 140.3, 145.0, 170.0, 172.1. MS: [M þ Na]þ calcd for
C16H24N2O3Na, 315.1685; found, 315.1673
N-(4-Ethylphenyl)-N0-hydroxyoctanediamide (7). White solid
(97%), mp 146 ꢀC. 1H NMR (DMSO-d6) δ (ppm): 1.14 (t, 3H,
CH2CH3), 1.26 (m, 4H, 2H4, and 2H5), 1.49 (m, 4H, 2H3, and
2H6), 1.93 (m, 2H, 2H7), 2.27 (t, 2H, 2H2), 2,51 (q, 2H,
CH2CH3), 7.10 (d, 2H, 2Ar-H), 7.49 (d, 2H, 2Ar-H), 8.65 (s,
1H, OH), 9.84 (s, 1H, NH), 10.35 (s, 1H, NH). 13C NMR
(DMSO-d6) δ (ppm): 21.3, 25.9, 29.3, 33.1, 37.2, 78.6, 119.9,
129.9, 132.6, 137.7, 169.9, 171.8. MS: [M þ Na]þ calcd for
C16H24N2O3Na, 315.1685; found, 315.1692
N-(2,3-Dimethylphenyl)-N0-hydroxyoctanediamide (8). White
solid (50%), mp 129 ꢀC. 1H NMR (DMSO-d6) δ (ppm): 1.30 (m,
4H, 2H4, and 2H5), 1.54 (m, 4H, 2H3, and 2H6), 1.95 (t, 2H,
2H7), 2.05 (s, 3H, CH3), 2.24 (s, 3H, CH3), 2.30 (t, 2H, 2H2), 7.03
(m, 3H, 3Ar-H), 8.68 (s, 1H, OH), 9.36 (s, 1H, NH), 10.40 (s,
1H, NH). 13C NMR (DMSO-d6) δ (ppm): 15.0, 21.1, 26.0, 26.2,
29.3, 33.2, 36.6, 124.6, 126.6, 127.6, 132.1, 137.1, 137.7, 170.2,
172.1. MS: [M þ Na]þ calcd for C16H24N2O3Na, 315.1685;
found, 315.1685
SAHA. White solid, mp 160 ꢀC. 1H NMR (DMSO-d6) δ
(ppm): 1.27 (m, 4H, 2H4, and 2H5), 1.52 (m, 4H, 2H3, and 2H6),
1.94 (m, 2H, 2H7), 2.29 (m, 2H, 2H2), 7.00 (m, 1H, Ar-H), 7.27
(m, 2H, 2Ar-H), 7.59 (d, 2H, 2Ar-H), 8.66 (s, 1H, OH), 9.92 (s,
1H, NH), 10.29 (s, 1H, NH).
N-Hydroxy-N0-(4-methoxyphenyl)octanediamide (1). White
1
N-(2,4-Dimethylphenyl)-N0-hydroxyoctanediamide (9). White
solid (57%), mp 117 ꢀC. 1H NMR (DMSO-d6) δ (ppm): 1.30 (m,
4H, 2H4, and 2H5), 1.53 (m, 4H, 2H3, and 2H6), 1.95 (t, 2H,
2H7), 2.13 (s, 3H, CH3), 2.24 (s, 3H, CH3), 2.30 (t, 2H, 2H2), 6.94
(d, 1H, 1Ar-H), 6.99 (s, 1H, 1Ar-H), 7.19 (d, 1H, 1Ar-H),
8.67 (s, 1H, OH), 9.18 (s, 1H, NH), 10.37 (s, 1H, NH). 13C NMR
(DMSO-d6) δ (ppm): 18.7, 21.6, 26.0, 26.2, 29.3, 33.1, 32.6, 39.1,
126.2, 127.2, 131.6, 132.7, 134.7, 134.9, 135.1, 170.0, 172.0. MS:
[M þ Na]þ calcd for C16H24N2O3Na, 315.1685; found, 315.1688
N-(2,5-Dimethylphenyl)-N0-hydroxyoctanediamide (10). White
solid (51%), mp 151 ꢀC. H NMR (DMSO-d6) δ (ppm): 1.26
(m, 4H, 2H4, and 2H5), 1.50 (m, 4H, 2H3, and 2H6), 1.93 (m, 2H,
2H7), 2.25 (m, 2H, 2H2), 3.70 (m, 3H, OMe), 6.83 (d, 2H,
2Ar-H), 7.48 (d, 2H, 2Ar-H), 8.66 (s, 1H, OH), 9.78 (s, 1H,
NH), 10.37 (s, 1H, NH). 13C NMR (DMSO-d6) δ (ppm): 25.9,
26.0, 29.3, 33.1, 37.1, 56.0, 114.6, 121.4, 133.5, 155.8, 170.0,
171.6. MS: [M þ Na]þ calcd for C15H22N2O4Na, 317.1477;
found, 317.1477.
N-Hydroxy-N0-(2-methylphenyl)octanediamide (2). White so-
lid (38%), mp 135 ꢀC. 1H NMR (DMSO-d6) δ (ppm): 1.31 (m,
4H, 2H4, and 2H5), 1.54 (m, 4H, 2H3, and 2H6), 1.95 (t, 2H,
2H7), 2.18 (s, 3H, CH3), 2.32 (m, 5H, 2H2, and CH3), 7.00-7.30
(m, 4H, 4Ar-H), 8.67 (s, 1H, OH), 9.23 (s, 1H, NH), 10.35 (s,
1H, NH). 13C NMR (DMSO-d6) δ (ppm): 18.7, 25.9, 26.1, 29.3,
33.1, 33.6, 36.6, 125.9, 126.1, 126.7, 131.1, 132.7, 137.3, 170.0,
172.0. MS: [M þ Na]þ calcd for C15H22N2O3Na, 301.1528;
found, 301.1525.
N-Hydroxy-N0-(3-methylphenyl)octanediamide (3). White so-
lid (52%), mp 118 ꢀC. 1H NMR (DMSO-d6) δ (ppm): 1.26 (m,
4H, 2H4, and 2H5), 1.48 (m, 2H, 2H6), 1.56 (m, 2H, 2H3), 1.93 (t,
2H, 2H7), 2.25 (m, 5H, 2H2, and CH3), 6.83 (d, 1H, Ar-H), 7.14
(t, 1H, Ar-H), 7.36 (d, 1H, Ar-H), 7.43 (s, 1H, Ar-H), 8.66 (s,
1H, OH), 9.79 (s, 1H, NH), 10.33 (s, 1H, NH). 13C NMR
(DMSO-d6) δ (ppm): 22.1, 25.9, 29.2, 33.1, 37.2, 117.1, 120.4,
124.5, 129.3, 138.6, 140.1, 169.9, 172.1. MS: [M þ Na]þ calcd for
C15H22N2O3Na, 301.1528; found, 301.1523
1
solid (52%), mp 134 ꢀC. H NMR (DMSO-d6) δ (ppm): 1.30
(m, 4H, 2H4, and 2H5), 1.54 (m, 4H, 2H3, and 2H6), 1.95 (t, 2H,
2H7), 2.12 (s, 3H, CH3), 2.24 (s, 3H, CH3), 2.30 (t, 2H, 2H2), 6.87
(d, 1H, 1Ar-H), 7.05 (d, 1H, 1Ar-H), 7.17 (s, 1H, 1Ar-H), 8.67
(s, 1H, OH), 9.20 (s, 1H, NH), 10.37 (s, 1H, NH). 13C NMR
(DMSO-d6) δ (ppm): 18.4, 21.5, 26.0, 26.2, 29.3, 33.1, 33.7, 122.5,
126.6, 129.5, 130.9, 135.7, 137.2, 170.0, 172.0. MS: [M þ Na]þ
calcd for C16H24N2O3Na, 315.1685; found, 315.1678
N-(2,6-Dimethylphenyl)-N0-hydroxyoctanediamide (11). White
solid (95%), mp 133 ꢀC. 1H NMR (DMSO-d6) δ (ppm): 1.31 (m,
4H, 2H4, and 2H5), 1.50 (m, 4H, 2H3, and 2H6), 1.95 (t, 2H, 2H7),
2.12 (s, 6H, CH3), 2.30 (t, 2H, 2H2), 7.05 (s, 3H, 3Ar-H), 8.81 (s,
1H, OH), 9.23 (s, 1H, NH), 10.39 (s, 1H, NH). 13CNMR(DMSO-
d6) δ (ppm): 19.0, 26.0, 26.3, 29.3, 29.4, 33.1, 36.2, 127.1, 128.5,
136.0, 136.3, 170.0, 171.7. MS: [M þ Na]þ calcd for C16H24N2O3-
Na, 315.1685; found, 315.1677
N-(2-Ethylphenyl)-N0-hydroxyoctanediamide (5). White solid
(71%), mp 116 ꢀC. 1H NMR (DMSO-d6) δ (ppm): 1.10 (t, 3H,
CH2CH3), 1.29 (m, 4H, 2H4, and 2H5), 1.53 (m, 4H, 2H3, and
2H6), 1.95 (t, 2H, 2H7), 2.31 (t, 2H, 2H2), 2.51 (q, 2H, CH2CH3),
7.21 (m, 4H, 4Ar-H), 8.67 (s, 1H, OH), 9.30 (s, 1H, NH), 10.39
(s, 1H, NH). 13C NMR (DMSO-d6) δ (ppm): 15.2, 24.7, 25.9,
N-(3,4-Dimethylphenyl)-N0-hydroxyoctanediamide (12). White
solid (42%), mp 134 ꢀC. 1H NMR (DMSO-d6) δ (ppm): 1.26 (m,
4H, 2H4, and t 2H5), 1.53 (m, 4H, 2H3, and 2H6), 2.16 (m, 10H,
2CH3, 2H2, and 2H7), 7.00 (d, 1H, Ar-H), 7.28 (d, 1H, Ar-H),
7.35(s, 1H, Ar-H), 8.66 (s, 1H, OH), 9.71 (s, 1H, NH), 10.35 (s,
1H, NH). 13C NMR (DMSO-d6) δ (ppm): 19.6, 20.5, 23.9, 26.0,

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 5 1947
29.3, 33.1, 117.5, 121.2, 130.3, 131.4, 137.0, 138.0, 170.0, 171.9.
MS: [M þ Na]þ calcd for C16H24N2O3Na, 315.1685; found,
315.1673
(s, 3H, CH3), 2.17 (t, 2H, 2H7), 2.24 (s, 3H, CH3), 2.28 (t, 2H,
2H2), 6.94 (d, 1H, 1Ar-H), 7.00 (s, 1H, 1Ar-H), 7.20 (d, 1H,
Ar-H), 9.16 (s, 1H, NH). 13C NMR (DMSO-d6) δ (ppm): 18.7,
21.3, 25.3, 26.1, 26.2, 29.2, 29.3 29.4, 34.5, 36.5, 126.2, 127.2,
131.5, 132.6, 134.7, 135.0, 172.0, 175.4. MS: Mþ calcd for
C16H23NO3, 277.1678; found, 277.1686
8-[(2,5-Dimethylphenyl)amino]-8-oxooctanoic Acid (10a). White
solid (48%), mp 128 ꢀC. 1H NMR (DMSO-d6) δ (ppm): 1.31 (m,
4H, 2H4, and 2H5), 1.54 (m, 4H, 2H3, and 2H6), 2.13 (m, 5H, 2H7,
and CH3), 2.22 (m, 5H, 2H2, and CH3), 6.88 (d, 1H, Ar-H), 7.06
(d, 1H, Ar-H), 7.17 (s, 1H, 1Ar-H), 9.17 (s, 1H, NH). 13C NMR
(DMSO-d6) δ (ppm): 18.3, 21.5, 25.3, 26.1, 29.2, 29.3, 34.6, 36.6,
126.6, 129.5, 130.9, 135.7, 137.1, 172.0, 175.4. MS: Mþ calcd for
C16H23NO3, 277.1678; found, 277.1686
8-[(2,6-Dimethylphenyl)amino]-8-oxooctanoic Acid (11a).
White solid (22%), mp 129 ꢀC. 1H NMR (DMSO-d6) δ (ppm):
1.32 (m, 4H, 2H4, and 2H5), 1.55 (m, 4H, 2H3, and 2H6), 2.12 (s,
6H, CH3), 2.21 (t, 2H, 2H7), 2.30 (t, 2H, 2H2), 7.05 (s, 3H,
3Ar-H), 9.18 (s, 1H, NH). 13C NMR (DMSO-d6) δ (ppm): 19.1,
25.3, 26.2, 29.2, 29.4, 34.5, 36.2, 127.1, 128.5, 136.0, 136.3, 171.6,
175.4. MS: Mþ calcd for C16H23NO3, 277.1678; found, 277.1658
8-[(3,4-Dimethylphenyl)amino]-8-oxooctanoic Acid (12a).
White solid (56%), mp 120 ꢀC, 1H NMR (DMSO-d6) δ (ppm):
1.29 (m, 4H, 2H4, and 2H5), 1.53 (m, 4H, 2H3, and 2H6), 2.22 (m,
10H, 2 CH3-Ar, 2H2, and 2H7), 7.00 (d, 1H, Ar-H), 7.32 (d,
1H, Ar-H), 7.36 (s, 2H, 2Ar-H), 9.69 (s, 1H, NH). 13C NMR
(DMSO-d6) δ (ppm): 19.6, 20.5, 25.3, 25.9, 29.3, 34.5, 37.2,
117.5, 121.2, 130.3, 131.4, 137.0, 138.0, 171.8, 175.4. MS: Mþ
calcd for C16H23NO3, 277.1678; found, 277.1686
N-(3,5-Dimethylphenyl)-N0-hydroxyoctanediamide (13). White
solid (53%), mp 171 ꢀC. 1H NMR (DMSO-d6) δ (ppm): 1.27 (m,
4H, 2H4, and 2H5), 1.51 (m, 4H, 2H3, and 2H6), 1.94 (t, 2H, 2H7),
2.21 (s, 6H, CH3), 2.27 (t, 2H, 2H2), 6.65 (s, 1H, 1Ar-H), 7.22 (s,
2H, 2 Ar-H), 8.67 (s, 1H, OH), 9.76 (s, 1H, NH), 10.35 (s, 1H,
NH). 13C NMR (DMSO-d6) δ (ppm): 22.0, 25.9, 26.0, 29.3, 33.1,
37.2, 117.7, 125.3, 138.4, 140.1, 170.0, 172.0. MS: [M þ Na]þ calcd
for C16H24N2O3Na, 315.1685; found, 315.1677
8-[(4-Methoxyphenyl)amino]-8-oxooctanoic Acid (1a). White
solid (21%), mp 144 ꢀC. 1H NMR (DMSO-d6) δ (ppm): 1.29 (m,
4H, 2H4, and 2H5), 1.51 (m, 4H, 2H3, and 2H6), 2.22 (m, 4H,
2H2, and 2H7), 3.70 (s, 3H, OCH3), 6.84 (m, 2H, 2Ar-H), 7.47
(m, 2H, 2Ar-H), 9.70 (s, 1H, NH). 13C NMR (DMSO-d6)
δ (ppm): 25.3, 25.9, 29.2, 29.3, 34.5, 37.1, 56.0, 114.6, 121.4,
133.4, 155.8, 171.5, 175.4. MS: [M þ Na]þ calcd for C15H21NO4,
279.1471; found, 279.1446
8-[(2-Methylphenyl)amino]-8-oxooctanoic Acid (2a). White
1
solid (41%), mp 126 ꢀC. H NMR (DMSO-d6) δ (ppm): 1.31
(m, 4H, 2H4, and 2H5), 1.53 (m, 4H, 2H3, and 2H6), 2.00-2.35
(m, 7H, 2H2, 2H7, and CH3), 6.95-7.40 (m, 4H, 4Ar-H), 9.24
(s, 1H, NH). 13C NMR (DMSO-d6) δ (ppm): 18.7, 25.2, 26.1,
29.2, 34.5, 36.6, 126.0, 126.2, 126.7, 131.1, 132.7, 137.3, 172.1,
175.4. MS: Mþ calcd for C15H21NO3Na, 286.1419; found,
286.1418
8-[(3-Methylphenyl)amino]-8-oxooctanoic Acid (3a). White
1
solid (26%), mp 102 ꢀC. H NMR (DMSO-d6) δ (ppm): 1.29
8-[(3,5-Dimethylphenyl)amino]-8-oxooctanoic Acid (13a).
White solid (30%), mp 123 ꢀC. 1H NMR (DMSO-d6) δ (ppm):
1.28 (m, 4H, 2H4, and 2H5), 1.53 (m, 4H, 2H3, and 2H6), 2.22 (m,
10H, 2H2, 2H7, and CH3), 6.66 (s, 1H, 1Ar-H), 7.18 (s, 2H,
2Ar-H), 9.69 (s, 1H, NH). 13C NMR (DMSO-d6) δ (ppm): 22.0,
25.3, 26.0, 29.3, 34.5, 37.2, 117.7, 125.3, 138.4, 140.1, 172.0,
175.4. MS: Mþ calcd for C16H23NO3, 277.1678; found, 277.1686
Cell Lines and Drug Treatment. All compounds (1-13, in-
cluding SAHA) were dissolved in DMSO (dimethyl sulfoxide,
Sigma) at 10 mM as a stock solution and stored at -20 ꢀC. The
Caco-2 cell line (ATCC number HTB-37), derived from a
human colorectal adenocarcinoma, as well as HepG2 cell line
(ATCC number HB-8065) were maintained under the condi-
tions recommended by the suppliers at 37 ꢀC and 5% CO2.
MUF is a primary cell line of human foreskin diploid fibroblasts
(m, 4H, 2H4, and 2H5), 1.53 (m, 4H, 2H3, and 2H6), 2.25 (m, 7H,
2H2, 2H7, and CH3), 6.83 (d, 1H, Ar-H), 7.15 (t, 1H, Ar-H),
7.36 (d, 1H, Ar-H), 7.43 (s, 1H, Ar-H), 9.75 (s, 1H, NH). 13C
NMR (DMSO-d6) δ (ppm): 22.1, 25.3, 25.9, 29.2, 29.3, 34.5,
37.2, 117.1, 120.4, 124.5, 129.3, 138.7, 140.1, 172.1, 175.4. MS:
Mþ calcd for C15H21NO3, 263.1521; found, 263.1510
8-[(2-Ethylphenyl)amino]-8-oxooctanoic Acid (5a). White so-
1
lid (26%), mp 102 ꢀC. H NMR (DMSO-d6) δ (ppm): 1.11 (t,
3H, CH2CH3), 1.31 (m, 4H, 2H4, and 2H5), 1.55 (m, 4H, 2H3,
and 2H6), 2.21 (t, 2H, 2H7), 2.31 (t, 2H, 2H2), 2.50 (q, 2H,
CH2CH3), 7.22 (m, 4H, 1Ar-H), 9.24 (s, 1H, NH). 13C NMR
(DMSO-d6) δ (ppm): 16.5, 26.1, 26.7, 27.5, 30.6, 35.9, 37.9,
127.8, 128.1, 128.5, 130.7, 138.0, 140.3, 173.7, 176.8. MS: Mþ
calcd for C16H23NO3, 277.1678; found, 277.1686
8-[(3-Ethylphenyl)amino]-8-oxooctanoic Acid (6a). White so-
1
ꢁ
created by Dr. Pirsel (Cancer Research Institute, Slovak Aca-
lid (38%), mp 102 ꢀC. H NMR (DMSO-d6) δ (ppm): 1.16 (t,
demy of Sciences, Slovak Republic). Human fibroblasts BJ
(ATCC number CRL-2522) were established from skin taken
from normal foreskin. MUF and BJ cell lines were normal and
nontransformed and, therefore, suitable as controls to tumor
cell lines. MUF and BJ cell lines were cultured in DMEM
supplemented with antibiotics and 10% fetal calf serum. For
protein extract preparation and Western blotting analysis, cells
were plated 1 day before drug treatment in six-well plates at a
density of 3 ꢀ 105 cells per well. Cells were treated with desired
compound at concentrations and times indicated in the figure
captions.
Ex Vivo HDAC Activity. All compounds were evaluated for
their ability to inhibit total HDAC activity ex vivo in Caco-2
cells using the HDAC Fluorimetric Cellular Activity Assay Kit
(Biomol International) and employing TSA as a positive con-
trol. This assay is based on the use of Fluor de Lys fluorescent
substrate, which is cell-permeable and a substrate for most class
1 and 2 HDACs. Caco-2 cells were seeded in triplicate in 96-well
plates at a density of 3 ꢀ 104 cells per well and allowed to attach
overnight (70% confluency). Media were replaced with 50 μL
per well of media containing 200 μM Fluor de Lys substrate
and 1 μM TSA as a positive control. Plates were incubated at
37 ꢀC and 5% CO2 for 4 h. To terminate HDAC activity and
begin development of the fluorescence signal, 50 μL per well of
the 1ꢀ Developer (Biomol International) diluted in cold lysis
3H, CH2CH3), 1.19 (m, 4H, 2H4, and 2H5), 1.54 (m, 4H, 2H3,
and 2H6), 2.23 (m, 4H, 2H2, and 2H7), 2.54 (q, 2H, CH2CH3),
6.86 (d, 1H, Ar-H), 7.18 (t, 1H, Ar-H), 7.41 (d, 1H, Ar-H),
7.44 (s, 1H, Ar-H), 9.79 (s, 1H, NH). 13C NMR (DMSO-d6)
δ (ppm): 16.4, 25.3, 25.9, 29.2, 29.3, 34.5, 37.2, 117.3, 119.3,
123.3, 124.5, 129.4, 140.2, 145.0, 172.1, 175.4. MS: Mþ calcd for
C16H23NO3, 277.1678; found, 277.1686
8-[(4-Ethylphenyl)amino]-8-oxooctanoic Acid (7a). White so-
1
lid (19%), mp 138 ꢀC. H NMR (DMSO-d6) δ (ppm): 1.15 (t,
3H, CH2CH3), 1.29 (m, 4H, 2H4, and 2H5), 1.53 (m, 4H, 2H3,
and 2H6), 2.30 (m, 4H, 2H2, and 2H7), 2.53 (q, 2H, CH2CH3),
7.11 (d, 2H, 2Ar-H), 7.48 (d, 2H, 2Ar-H), 9.77 (s, 1H, NH).
13C NMR (DMSO-d6) δ (ppm): 16.7, 25.3, 25.9, 29.2, 34.5, 37.2,
120.0, 128.7, 137.9, 139.1, 171.8, 175.4. MS: Mþ calcd for
C16H23NO3, 277.1678; found, 277.1686
8-[(2,3-Dimethylphenyl)amino]-8-oxooctanoic Acid (8a). White
solid (42%), mp 142 ꢀC. 1H NMR (DMSO-d6) δ (ppm): 1.32 (m,
4H, 2H4, and 2H5), 1.55 (m, 4H, 2H3, and 2H6), 2.05 (t, 3H, CH3),
2.24 (m, 7H, 2H2, 2H7, and CH3), 7.03 (m, 3H, 3Ar-H), 9.28 (s,
1H, NH). 13C NMR (DMSO-d6) δ (ppm): 14.9, 21.1, 25.3, 26.1,
29.2, 29.3, 34.6, 36.6, 124.6, 126.0, 127.7, 132.1, 137.1, 137.7, 172.0,
175.4. MS: Mþ calcd for C16H23NO3, 277.1678; found, 277.1658
8-[(2,4-Dimethylphenyl)amino]-8-oxooctanoic Acid (9a).
White solid (32%), mp 122 ꢀC. 1H NMR (DMSO-d6) δ (ppm):
1.31 (m, 4H, 2H4, and 2H5), 1.51 (m, 4H, 2H3, and 2H6), 2.13

1948 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 5
Oger et al.
buffer [50 mM Tris/HCl (pH 8.0), 137 mM NaCl, 2.7 mM KCl,
1 mM MgCl2, and 1% NP-40] þ 2 μM TSA was added and
mixed by pipetting. After developer was added, plates were
incubated for an additional 15 min at 37 ꢀC, and the fluore-
scence was read (excitation 365 nm and emission 450 nm) in
a microplate-reading fluorimeter (Fluorolite 1000, Dynatech
Laboratories). Comparison of inhibited vs control relative
fluorescence was employed to determine the percent HDAC
activity remaining.
chromatography. HDAC enzymes were diluted in HDAC buf-
fer in the presence of 0.01% (w/v) BSA (Sigma) at the following
working concentrations (1ꢀ): 75 nM HDAC1, 1 nM HDAC3,
1 nM HDAC5, 3 nM HDAC6, and 0.5 nM HDAC7. The
respective enzyme concentrations were optimized to ensure the
detection of the HDAC activities in the linear range of the
reaction for at least 60 min, with a consumption of less than 10%
of the initial substrate. Specific fluorogenic substrates were used
at working concentrations equivalent to their respective 2 ꢀ Km
values. In particular, the class 1 HDAC substrate (#50032) was
used at 12 μM (HDAC1 and HDAC3), the class 2a HDAC
substrate (#50040) was used at 4 μM (HDAC5 and HDAC7),
while the specific HDAC6 substrate (#50037) was used at
30 μM. Compounds dissolved in DMSO were assayed in quad-
ruplicate at 11 concentrations ranging from 0.316 nM to
31.6 μM for HDAC1, HDAC3, and HDAC6 and from
6.32 nM to 632 μM for HDAC5 and HDAC7. All of the dilution
steps were robotically performed with a Matrix Control Mate
(Thermo Scientific, Hudson, NH). The reaction was incubated
for 45 min at 32 ꢀC and stopped with developer 2ꢀ in the
presence of 8 μM TSA (Sigma). Plates were further incubated at
32 ꢀC for 30 min. The fluorescence intensity was detected with
Safire2 (Tecan, Mannedorf, Switzerland) at gain 40 and with
excitation-emission wavelengths of 353-450 nm, respectively.
End point relative fluorescence units (RFU) data were trans-
formed as a percentage of activity to compare the relative
potency of the compounds on the different HDAC enzymes.
Curve fitting was performed by GraphPad Prism 4.0, using a
sigmoidal dose-response (variable slope) model. The average
IC50 values were obtained as a minimum from two indepen-
dent experiments (for details, see the Supporting Information,
S4-S6).
Indirect ELISA. Whole cell extracts were prepared using a
Nuclear Extract Kit (Active Motif) according the manufac-
turer’s instructions using TSA (1 μM) in each step to avoid
protein deacetylation. Protein extracts (antigen) were diluted to
a final concentration of 20 μg/mL in PBS, and wells of a PVC
microtiter plate were coated with 1 μg of antigen. Plates were
covered and incubated for 2 h at room temperature. Then, after
the coating solution was removed, plates were washed three
times by filling the wells with 200 μL of PBS. The solutions or
washes were removed by flicking the plate over a sink. The
remaining protein binding sites in the coated wells were blocked
by adding 200 μL of blocking buffer [5% bovine serum albumin
(BSA) in PBS] per well. Plates were covered and incubated for
2 h at room temperature. Plates were then washed twice with
PBS, and 100 μL of antiacetyl lysines antibody (from rabbit,
diluted 1/1000, Abcam) diluted in blocking buffer was added to
each well. Plates were covered and incubated overnight at 4 ꢀC.
Subsequently, plates were washed four times with PBS, and
100 μL of conjugated secondary antibody (antirabbit, diluted
1/2500, Abcam) diluted in blocking buffer was added to
each well. Then, plates were covered and incubated 2 h at
room temperature. After four washes, the horseradish per-
oxidase (HRP) activity was measured using the colorimetric kit
(Substrate Reagent Pack, R&D Systems) following the supplier’s
recommendations.
Protein Expression and Purification of HDAC7. The catalytic
domain of human HDAC7 (Uniprot: Q8WUI4; amino acids
483-903 with an expression-enhancing NASNNG C-terminal
sequence) was expressed from a pET9-derived vector encoding
an N-terminal TEV cleavable hexahistidine tag. Escherichia coli
BL21 AI (Invitrogen) cotransformed with RIL plasmid
(Stratagene) was used for protein expression. Protein was
purified as described19 with minor modifications (Supporting
Information, S13).
Western Blot Analysis. After the media were removed from
six-well plates, cells were washed with PBS and harvested in
150 μL of PBS using a scraper. Protein concentrations were
directly quantified in cell suspension using DC Protein Assay
Kit (Biorad) according to the manufacturer’s instructions. After
centrifugation (5 min, 200g, 4 ꢀC), cell pellets were resuspended in
€
Laemmli loading buffer to obtain the desired protein concentra-
tion. Then, samples were denatured by sonication (bioruptor,
Diagenode, 7 min, 4 ꢀC, max power) and boiling (10 min, 100 ꢀC).
In all Western blotting analyses, equal protein quantities were
loaded on 12% SDS-PAGE, and electrophoresis separation was
followed by proteins transfer on nitrocellulose membranes
(Hybond-C Extra, Amersham Biosciences). The quality of protein
transfer was systematically verified using red Ponceau staining
(data not shown). Membranes were probed with appropriate
primary antibodies in milk saturation buffer (PBS/0.1% Tween,
5% milk) for anti-p21waf1/cip1 (from mouse, diluted 1/1000, Milli-
pore), antiactin (from mouse, diluted 1/5000, Abcam), antiacety-
lated H3 (from rabbit, diluted 1/25000, Abcam), antihistone H3
(from rabbit, diluted 1/2000, Abcam), antiacetylated R-tubulin
(from mouse, diluted 1/1000, Abcam), and anti-R-tubulin (from
mouse, diluted 1/1000, Sigma) or in BSA saturation buffer (PBS/
0.1% Tween, 5% BSA) for antiacetyl lysine (from rabbit, diluted
1/1000, Abcam). The secondary antibodies corresponding to
antimouse (diluted 1/10000, Santa Cruz) or antirabbit (diluted
1/2500, Abcam) were coupled to HRP. The signal was enhanced
by chemiluminescence using Immobilon (Millipore) as a substrate
for HRP and detected on film (Biomax Film, Kodak) or camera
(Biorad).
Acquisition of NMR Spectra. All NMR spectra were recorded
at 298 K with a spectral width of 14 ppm on Bruker Avance
500 MHz spectrometer, equipped with a 5 mm TXI inverse
triple-resonance cryoprobe head. (PRISM platform, University
of Rennes 1, France). Spectra processing was performed using
TOPSPIN 1.3 software (Bruker). The in vitro experiments were
performed as already described.34
STD Experiments. Selective saturation of the protein was
achieved by a train of Gauss-shaped pulses of 50 ms length each,
truncated at 1%, and separated by a 1 ms delay. Forty selective
pulses were applied, leading to a total length of the satura-
tion train of 2 s. The on-resonance irradiation of the protein
was performed at a chemical shift of 1.3 ppm. Off-resonance
irradiation was set at 40 ppm, where no protein signals were
present. Unwanted magnetization water was suppressed using a
WATERGATE 3-9-19 pulse sequence. A 30 ms spin lock pulse
was applied to remove protein background signals. The total
scan number in the STD experiments was 2000, acquired as 16
blocks of 128 scans each. The spectra on- and off-resonances
were subtracted after multiplication by an exponential line-
broadening function of 0.8 Hz prior to Fourier transformation.
WaterLOGSY Experiments. WaterLOGSY experiments
were performed using a selective excitation with a 180ꢀ shaped
pulse at H2O position. The first water selective 180ꢀ pulse was
applied during 5 ms. The first two pulsed field gradients (PFGs)
had a typical duration of 2 ms. A weak rectangular PFG is
applied during the entire length of the mixing time (1 s). A short
gradient recovery time of 1 ms is applied at the end of the
mixing time before the detection pulse. The water suppression in
Compound Profiling on HDAC1, HDAC3, HDAC5, HDAC6,
and HDAC7. HDAC1, HDAC3, and HDAC6 enzymes, HDAC
substrates, HDAC buffer, and developer were from BPS
Bioscience (San Diego, CA). Human HDAC5 and HDAC7
were recombinantly expressed in Sf9þ insect cells as chimeric
fusions with N-terminal GST and C-terminal HIS tag and
then purified by glutathione-sepharose and Ni-NTA Superflow

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 5 1949
both experiments was achieved with the excitation sculpt-
ing sequence. The data were collected with a sweep width
of 14 ppm, an acquisition time of 0.585 s, and a relaxation
delay of 2 s. Prior to Fourier transformation, the data were
multiplied with an exponential function with a line-broadening
of 0.8 Hz.
background was measured at 630 nm using a microplate
reader (Power Wave X5; Biotek). Data were acquired using
KC4 version 3.4 software (Biotek Instruments Inc.). Results
(% of cell growth) were calculated as follows: [(A570
630 compound/(A570 - A630 DMSO
from three independent experiments for Caco-2 cells and from at
least six independent experiments for BJ and MUF cells. Curve
fitting was performed by GraphPad Prism 4.0.
-
A
)
)
] ꢀ 100. Results were averaged
Molecular Modeling. Molecular modeling studies were per-
formed with Macromodel 835 on a Intel Xeon dual core with
Fedora Core 7 as the operating system. The OPLSAA-2001
force field35,36 was applied for energy minimization and mole-
cular dynamics simulation. Energy minimization protein-
ligand complexes were performed with the conjugate gradient
€
Acknowledgment. We thank Maud Bizot and Gaelle
Palierne for technical assistance. We thank Jean-Franc-ois
Geffroy, Anne-Sophie Julien, Alicia Cassant, and Elodie
ꢀ
Guiheux for their technical support. We thank Dr. Jerome
Eeckhoute for his precious help with the writing of the manu-
ꢁ
script. MUF cells were kindly provided by Dr. Miroslav Pirsel.
This work was funded by the European Community (LSHM-
method37 to an energy gradient of 0.001 kJ/(mol A) with distant-
˚
dependent electrostatic treatment. The three ligands SAHA, 11,
and 13 were built by the Build tool and energy minimized with
previous force field and parameters.
Protein-Ligand Complex Simulations. The crystals com-
plexes HDAC7-SAHA (PDB ID: 3C0Z) and HDAC7-TSA
(PDB ID: 3C10) have been downloaded from the public RCSB
Protein Data Bank (www.rcdb.org).38 The chains A of all
complexes were used for all of the simulations. In all three
complexes of HDAC7-SAHA, the ligands were not completely
resolved, and particularly, the N-phenylacetamide was not
present. This precludes evaluation of the interactions of the
N-phenylacetamide moiety with the amino acids of the protein,
and using the complex HDAC7-SAHA with a truncate ligand
in the preliminary minimization phases could bring partial
occlusion of the entrance of the cavity by the closest amino
acids side chains. After these considerations from both crystal
cells, the two A chains have been selected and superimposed
ꢀ
CT-2006-037498), the Region Bretagne (SIE 211-B3-11),
the Centre National pour la Recherche Scientifique, and the
ꢂ
Ministere de l’Enseignement Superieur et de la Recherche.
ꢀ
Supporting Information Available: Acetylation level of his-
tones H3/H4 and R-tubulin in response to SACs treatment (S1),
materials and methods for histone preparation (S2), Western
blot analysis of protein acetylation in HepG2 cells treated with
SACs (S3), compound handling and pipetting scheme for com-
pound profiling (S4), Z0 factor evaluation39 (S5), dose-reponse
curves on HDAC1, -3, -5, -6, and -7 using SACs compounds
(S6), modeling experiments suggesting a conserved pro-
tein-ligand interaction mode between the cap domain of SAHA
and class 2 HDACs, method for HDAC4-SACs complex
modeling (S7), importance of HDAC7 proline 809 and HDAC4
proline 274 for ligand interaction (S8), lengthening of the carbon
chain on the phenyl ring of SAHA leading to eratic docking (S9),
p21waf1/cip1 induction in response to increasing concentrations
(5, 50, and 100 μM) of representative o-substituted SACs (S10),
MTT assay and p21waf1/cip1 induction in HepG2 cells
treated with SACs (S11), table of RP-HPLC purity and micro-
analytical data (S12), and complete material and methods for
recombinant HDAC7 preparation and purification (S13). This
material is available free of charge via the Internet at http://
pubs.acs.org.
˚
(rmsd on the heavy atoms of the backbone was 0.134 A) to
evaluate the peculiar differences in the amino acid side chain
orientations or in the tertiary structure near the binding cavity
and in the receptor surface close to the entrance of the cavity.
After this checking phase, where no important differences were
found, the chain A of the complex HDAC7-TSA was selected
for the successive simulations. All hydrogens were inserted into
the structure of chain A of 3C10 complex, all water molecules
were removed, the NδH tautomer of the H709 was chosen, and a
partial charge of þ2 was assigned to the zinc ions. The chain A,
with TSA, was energetically minimized with OPLSAA-2001
force field keeping the entire backbone fixed. The TSA ligand
was removed from the minimized complex, and the truncated
SAHA has been introduced as a guideline for the successive
insertion of the complete SAHA, 11, and 13. The three com-
plexes HDAC7-SAHA, HDAC7-11, and HDAC7-13 were
generated, and their energies were minimized. All of the mini-
mum energy conformations of the complexes were submitted
to dynamic simulations with the OPLSAA-2001 force field
at 300 K, with an 10 ps of equilibration, and 100 ps of produc-
tion time with a time step of 1 fs collecting 100 snapshots every
References
(1) Yang, X. J.; Seto, E. The Rpd3/Hda1 family of lysine deacetylases:
from bacteria and yeast to mice and men. Nat. Rev. Mol. Cell Biol.
2008, 9, 206–218.
(2) Michan, S.; Sinclair, D. Sirtuins in mammals: Insights into their
biological function. Biochem. J. 2007, 404, 1–13.
(3) Lu, Q.; Qiu, X.; Hu, N.; Wen, H.; Su, Y.; Richardson, B. C.
Epigenetics, disease, and therapeutic interventions. Ageing Res.
Rev. 2006, 5, 449–467.
(4) Paris, M.; Porcelloni, M.; Binaschi, M.; Fattori, D. Histone
deacetylase inhibitors: from bench to clinic. J. Med. Chem. 2008,
51, 1505–1529.
(5) Bolden, J. E.; Peart, M. J.; Johnstone, R. W. Anticancer activities
of histone deacetylase inhibitors. Nat. Rev. Drug. Discovery 2006, 5,
769–784.
(6) Mai, A.; Massa, S.; Lavu, S.; Pezzi, R.; Simeoni, S.; Ragno, R.;
Mariotti, F. R.; Chiani, F.; Camilloni, G.; Sinclair, D. A. Design,
Synthesis, And Biological Evaluation of Sirtinol Analogues as
Class III Histone/Protein Deacetylase (Sirtuin) Inhibitors.
J. Med. Chem. 2005, 48, 7789–7895.
˚
1 ps. Only the amino acid side chains inside a sphere of 5 A from
the ligand were free to move, while the rest of the amino acid side
chains and the entire backbone were kept fixed. The average
structures obtained from 10 representative snapshots of MD
trajectory were energy minimized with the OPLSAA-2001 force
˚
field to a gradient of 0.001 kJ/(mol A), with a fixed protein
backbone.
Growth Inhibition Assays. One day prior to treatment,
Caco-2, HepG2, and normal human fibroblast cells (MUF
and BJ cells) were plated on 96-well plates in appropriate media,
at densities of 2.104 cells/well for Caco-2 and HepG2 cells and
4.103 cells/well for MUF and BJ cells. Cells were exposed to
the tested compounds for 48 h at concentrations ranging from
100 to 0.05 μM with 2-fold dilutions. DMSO was used as a
negative control with the same dilutions starting from 1%. After
treatment for 48 h, 10 μL of MTT (thiazolyl blue, 5 mg/mL) was
added to each well and cultured for 2 h. Then, after the media
were removed, formazan crystals were dissolved in 100 μL of
DMSO. The absorbance was measured at 570 nm, and the
(7) Marks, P. A.; Breslow, R. Dimethyl sulfoxide to vorinostat:
Development of this histone deacetylase inhibitor as an anticancer
drug. Nat. Biotechnol. 2007, 25, 84–90.
(8) Miller, T. A.; Witter, D. J.; Belvedere, S. Histone Deacetylase
Inhibitors. J. Med. Chem. 2003, 46, 5097–5116.
(9) Finnin, M. S.; Donigian, J. R.; Cohen, A.; Richon, V. M.; Rifkind,
R. A.; Marks, P. A.; Breslow, R.; Pavletich, N. P. Structures of a
histone deacetylase homologue bound to the TSA and SAHA
inhibitors. Nature 1999, 401, 188–193.
(10) Vannini, A.; Volpari, C.; Filocamo, G.; Casavola, E. C.; Brunetti,
M.; Renzoni, D.; Chakravarty, P.; Paolini, C.; De Francesco, R.;

1950 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 5
Oger et al.
Gallinari, P.; Steinkuhler, C.; Di Marco, S. Crystal structure of a
eukaryotic zinc-dependent histone deacetylase, human HDAC8,
complexed with a hydroxamic acid inhibitor. Proc. Natl. Acad. Sci.
U.S.A. 2004, 101, 15064–15069.
to proteins via magnetization transfer from bulk water. J. Biomol.
NMR 2000, 18, 65–68.
(24) Fielding, L. NMR methods for the determination of protein-ligand
dissociation constants. Prog. Nucl. Magn. Reson. Spectrosc. 2007,
51, 219–242.
(11) Monneret, C. Histone deacetylase inhibitors. Eur. J. Med. Chem.
2005, 40, 1–13.
(25) Pellecchia, M.; Bertini, I.; Cowburn, D.; Dalvit, C.; Giralt, E.;
Jahnke, W.; James, T. L.; Homans, S. W.; Kessler, H.; Luchinat,
C.; Meyer, B.; Oschkinat, H.; Peng, J.; Schwalbe, H.; Siegal, G.
Perspectives on NMR in drug discovery: A technique comes of age.
Nat. Rev. Drug Discovery 2008, 7, 738–745.
(26) Ocker, M.; Schneider-Stock, R. Histone deacetylase inhibitors:
signalling towards p21cip1/waf1. Int. J. Biochem. Cell Biol. 2007,
39, 1367–1374.
(27) Kelly, W. K.; Marks, P. A. Drug insight: Histone deacetylase
inhibitors-development of the new targeted anticancer agent sub-
eroylanilide hydroxamic acid. Nat. Clin. Pract. Oncol. 2005, 2, 150–
157.
(28) Dowling, D. P.; Gantt, S. L.; Gattis, S. G.; Fierke, C. A.; Chris-
tianson, D. W. Structural Studies of Human Histone Deacetylase 8
and Its Site-Specific Variants Complexed with Substrate and
Inhibitors. Biochemistry 2008, 47, 13554–13563.
(12) Andrianov, V.; Gailite, V.; Lola, D.; Loza, E.; Semenikhina, V.;
Kalvinsh, I.; Finn, P.; Petersen, K. D.; Ritchie, J. W.; Khan, N.;
Tumber, A.; Collins, L. S.; Vadlamudi, S. M.; Bjorkling, F.;
Sehested, M. Novel amide derivatives as inhibitors of histone
deacetylase: design, synthesis and SAR. Eur. J. Med. Chem. 2009,
44, 1067–1085.
(13) Estiu, G.; Greenberg, E.; Harrison, C. B.; Kwiatkowski, N. P.;
Mazitschek, R.; Bradner, J. E.; Wiest, O. Structural Origin of
Selectivity in Class II-Selective Histone Deacetylase Inhibitors.
J. Med. Chem. 2008, 51, 2898–2906.
(14) Mai, A.; Esposito, M.; Sbardella, G.; Massa, S. A new facile and
expeditious synthesis of N-hydroxy-N0-phenyloctanediamide, a
potent inducer of terminal cytodifferentiation. OPPI Briefs 2001,
33, 391–394.
(15) Codd, R. Coordination chemistry of hydroxamic acid. Coord.
Chem. Rev. 2008, 252, 1387–1408.
(29) Kahnberg, P.; Lucke, A. J.; Glenn, M. P.; Boyle, G. M.; Tyndall, J.
D. A.; Parsons, P. G.; Fairlie, D. P. SAHA from R-Aminosuberic
Acid. J. Med. Chem. 2006, 49, 7611–7622.
(30) Avalos, M.; Babiano, R.; Barneto, J. L.; Bravo, J. L.; Cintas, P.;
Jimenez, J. L.; Palacios, J. C. Can We Predict the Conformational
Preference of Amides. J. Org. Chem. 2001, 66, 7275–7282.
(31) Stewart, W. E.; Siddall, T. H., III. Nuclear Magnetic Resonance
Studies of Amides. Chem. Rev. 1970, 70, 517–551.
(32) Mitro, N.; Godio, C.; De Fabiani, E.; Scotti, E.; Galmozzi, A.;
Gilardi, F.; Caruso, D.; Vigil Chacon, A. B.; Crestani, M. Insights
in the regulation of cholesterol 7alpha-hydroxylase gene reveal a
target for modulating bile acid synthesis. Hepatology 2007, 46, 885–
897.
(33) Witt, O.; Deubzer, H. E.; Milde, T.; Oehme, I. HDAC family: What
are the cancer relevant targets? Cancer Lett. 2009, 277, 8–21.
(34) Le Guevel, R.; Oger, F.; Lecorgne, A.; Dudasova, Z.; Chevance, S.;
Bondon, A.; Barath, P.; Simonneaux, G.; Salbert, G. Identification of
small molecule regulators of the nuclear receptor HNF4alpha based
on naphthofuran scaffolds. Bioorg. Med. Chem. 2009, 17, 7021–7030.
(35) Mohamadi, F.; Richards, N. G. J.; Guida, W. C.; Liskamp, R.;
Lipton, M.; Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C.
An integrated software for modeling organic and bioorganic
molecules using molecular mechanics. J. Comput. Chem. 1990,
11, 440–467.
(16) Lea, M. A.; Ibeh, C.; Shah, N.; Moyer, M. P. Induction of
differentiation of colon cancer cells by combined inhibition of
kinases and histone deacetylase. Anticancer Res. 2007, 27, 741–748.
(17) Mariadason, J. M. HDACs and HDAC inhibitors in colon cancer.
Epigenetics 2008, 3, 28–37.
(18) Khan, N.; Jeffers, M.; Kumar, S.; Hackett, C.; Boldog, F.;
Khramtsov, N.; Qian, X.; Mills, E.; Berghs, S. C.; Carey, N.; Finn,
P. W.; Collins, L. S.; Tumber, A.; Ritchie, J. W.; Jensen, P. B.;
Lichenstein, H. S.; Sehested, M. Determination of the class and
isoform selectivity of small-molecule histone deacetylase inhibi-
tors. Biochem. J. 2008, 409, 581–589.
(19) Bottomley, M. J.; Lo Surdo, P.; Di Giovine, P.; Cirillo, A.;
Scarpelli, R.; Ferrigno, F.; Jones, P.; Neddermann, P.; De
Francesco, R.; Steinkuhler, C.; Gallinari, P.; Carfi, A. Structural
and functional analysis of the human HDAC4 catalytic domain
reveals a regulatory structural zinc-binding domain. J. Biol. Chem.
2008, 283, 26694–26704.
(20) Schuetz, A.; Min, J.; Allali-Hassani, A.; Schapira, M.; Shuen, M.;
Loppnau, P.; Mazitschek, R.; Kwiatkowski, N. P.; Lewis, T. A.;
Maglathin, R. L.; McLean, T. H.; Bochkarev, A.; Plotnikov, A. N.;
Vedadi, M.; Arrowsmith, C. H. Human HDAC7 harbors a class IIa
histone deacetylase-specific zinc binding motif and cryptic deace-
tylase activity. J. Biol. Chem. 2008, 283, 11355–11363.
(21) Somoza, J. R.; Skene, R. J.; Katz, B. A.; Mol, C.; Ho, J. D.;
Jennings, A. J.; Luong, C.; Arvai, A.; Buggy, J. J.; Chi, E.; Tang, J.;
Sang, B. C.; Verner, E.; Wynands, R.; Leahy, E. M.; Dougan, D.
R.; Snell, G.; Navre, M.; Knuth, M. W.; Swanson, R. V.; McRee,
D. E.; Tari, L. W. Structural snapshots of human HDAC8 provide
insights into the class I histone deacetylases. Structure 2004, 12,
1325–1334.
(36) Jorgensen, W. L.; Maxwell, D. S.; Tirado-Rives, J. Development
and Testing of the OPLS All-Atom Force Field on Conformational
Energetics and Properties of Organic Liquids. J. Am. Chem. Soc.
1996, 118, 11225–11236.
ꢂ
(37) Polak, E.; Ribiere, G. Note sur la convergence de directions
ꢀ
conjuguees. Rev. Fr. Informat. Rech. Operationelle 1969, 16, 35–43.
(38) Berman, H. M.; Westbrook, J.; Feng, Z.; Gilliland, G.; Bhat, T. N.;
Weissig, H.; Shindyalov, I. N.; Bourne, P. E. The Protein Data
Bank. Nucleic Acids Res. 2000, 28, 235–42.
(39) Zhang, J. H.; Chung, T. D.; Oldenburg, K. R. A Simple Statistical
Parameter for Use in Evaluation and Validation of High Through-
put Screening Assays. J. Biomol. Screening 1999, 4, 67–73.
(22) Meyer, B.; Peters, T. NMR spectroscopy techniques for screening
and identifying ligand binding to protein receptors. Angew. Chem.,
Int. Ed. Engl. 2003, 42, 864–890.
(23) Dalvit, C.; Pevarello, P.; Tato, M.; Veronesi, M.; Vulpetti, A.;
Sundstrom, M. Identification of compounds with binding affinity