Synthesis and biochemical analysis of 2,2,3,3,4,4,5,5,6,6,7,7-dodecafluoro- N-hydroxy-octanediamides as inhibitors of human histone deacetylases

Article abstract of DOI:10.1016/j.bmc.2011.11.041

Inhibition of human histone deacetylases (HDACs) has emerged as a novel concept in the chemotherapeutic treatment of cancer. Two chemical entities, SAHA (ZOLINZA, Merck) and romidepsin (Istodax, Celgene) have been recently approved by the FDA as first-in-class drugs against cutaneous T-cell lymphoma. Clinical use of these drugs revealed several side effects including gastro-intestinal symptoms, fatigue, thrombocytopenia, thrombosis. Romidepsin is associated with an yet unresolved cardiotoxicity issue. A general hypothesis for the diminishment of unwanted adverse effects and an improved therapeutical window suggests the development of more isotype selective inhibitors. In this study the first time HDAC inhibitors with perfluorinated spacers between the zinc chelating moiety and the aromatic capping group were synthesized and tested against representatives of HDAC classes I, IIa and IIb. Competitive binding assays and a combined approach by using blind docking and molecular dynamics support binding of the perfluorinated analogs of SAHA to the active site of the HDAC-like amidohydrolase from Bordetella/Alcaligenes and presumably also to human HDACs. In contrast to the alkyl spacer of SAHA and derivatives, the perfluorinated alkyl spacer seems to contribute to or facilitate the induction of selectivity for class II, particularly class IIa, HDACs even though the overall potency of the perfluorinated SAHA analogs in this study against human HDACs remained still rather moderate in the micromolar range.

Full text of DOI:10.1016/j.bmc.2011.11.041

Bioorganic & Medicinal Chemistry 20 (2012) 985–995
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Synthesis and biochemical analysis of 2,2,3,3,4,4,5,5,6,6,7,7-dodecafluoro-
N-hydroxy-octanediamides as inhibitors of human histone deacetylases
Leonhard M. Henkes a, Patricia Haus a, Felix Jäger b, Joachim Ludwig b, Franz-Josef Meyer-Almes a,
⇑
a Department of Chemical Engineering and Biotechnology, University of Applied Sciences, Schnittspahnstraße 12, 64287 Darmstadt, Germany
b Synchem OHG, Am Kies 2, 34587 Felsberg/Altenburg, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Inhibition of human histone deacetylases (HDACs) has emerged as a novel concept in the chemothera-
peutic treatment of cancer. Two chemical entities, SAHA (ZOLINZA, Merck) and romidepsin (Istodax, Cel-
gene) have been recently approved by the FDA as first-in-class drugs against cutaneous T-cell lymphoma.
Clinical use of these drugs revealed several side effects including gastro-intestinal symptoms, fatigue,
thrombocytopenia, thrombosis. Romidepsin is associated with an yet unresolved cardiotoxicity issue. A
general hypothesis for the diminishment of unwanted adverse effects and an improved therapeutical
window suggests the development of more isotype selective inhibitors. In this study the first time HDAC
inhibitors with perfluorinated spacers between the zinc chelating moiety and the aromatic capping group
were synthesized and tested against representatives of HDAC classes I, IIa and IIb. Competitive binding
assays and a combined approach by using blind docking and molecular dynamics support binding of
the perfluorinated analogs of SAHA to the active site of the HDAC-like amidohydrolase from Borde-
tella/Alcaligenes and presumably also to human HDACs. In contrast to the alkyl spacer of SAHA and deriv-
atives, the perfluorinated alkyl spacer seems to contribute to or facilitate the induction of selectivity for
class II, particularly class IIa, HDACs even though the overall potency of the perfluorinated SAHA analogs
in this study against human HDACs remained still rather moderate in the micromolar range.
Ó 2011 Elsevier Ltd. All rights reserved.
Received 22 September 2011
Revised 15 November 2011
Accepted 20 November 2011
Available online 30 November 2011
Keywords:
Histone deacetylase
Histone deacetylase inhibitors
SAHA
Vorinostat
Oncology
1. Introduction
cess non histone substrates.5 Due to their nuclear and cytosolic
presence, HDACs impact many fundamental cellular functions,
The interaction between histone octamers and nuclear DNA de-
pends on the acetylation state of their N-terminal lysine residues.
By binding tightly to the DNA backbone, deacetylased lysine resi-
dues suppress the process of transcription, while acetylated lysine
residues support a further relaxed state, associated with increased
transcription rates.1 The process of deacetylation is catalyzed by a
group of proteins, the so called histone deacetylases (HDACs). In
contrast, histone acetyltransferases (HAT) reverse this process
antagonistically. HDACs are divided into four different classes
(I, II, III and IV) depending on their yeast homologs RPD3, HDA1
and Sir2.2,3 The classic HDACs share a coordinated Zn2+ ion in their
active site and comprise members of class I (HDACs 1, 2, 3 and 8),
class II (HDACs 4, 5, 6, 7, 9, 10) and class IV (HDAC11). Class II can
be subdivided into the nucleus–cytosol shuttling class IIa HDACs 4,
5, 7, 9 and the mainly cytoplasmic class IIb HDACs HDAC6 and
HDAC10, which posses two catalytic domains.4 Class III members,
the so called sirtuins, are distinguished from the classic HDACs
by their NAD+ dependent hydrolytic activity.2 In the cytosol class
II HDACs are involved in posttranslational modifications and pro-
including oncogenic and apoptotic pathways.6 In particular, differ-
ent HDAC isoforms contribute to numerous cell and tissue specific
processes during cancer induction and growth. The catalytic do-
main of Zn2+ containing HDACs is highly conserved and consists
of a tunnel leading to a buried Zn2+ ion within the active site, that
coordinates the substrate and a hydrolytic water molecule.7 The
reaction mechanism for classes I and IIb HDACs is supposed to be
mediated through
a histidine–aspartate charge-relay system
which increases the nucleophilic character of a Zn2+ coordinated
water molecule thereby enabling its nucleophilic attack on the
carbonyl carbon atom of the substrate. The resulting intermediate
is stabilized by an adjacent tyrosine residue. Then the carbon–
nitrogen bond is cleaved supported by a second charge-relay
system finally resulting in an acetate molecule and a deacetylated
lysine residue. Common histone deacetylase inhibitors (HDACi)
sharing the zinc chelating hydroxamate group like suberoylanilide
hydroxamic acid (SAHA) and Trichostatin A (TSA) intervene sub-
strate deacetylation by forming chelate complexes with the cata-
lytic Zn2+ ion. SAHA and TSA have been shown to interfere with
cell growth and prevent tumor differentiation in mice.6 However,
treatment of cancer patients with SAHA is accompanied with sev-
eral side effects which may result from poor isoform selectivity.
⇑
Corresponding author. Tel.: +49 (0) 6151168406; fax: +49 (0) 6151168404.
E-mail address: franz-josef.meyer-almes@h-da.de (F.-J. Meyer-Almes).
0968-0896/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2011.11.041

986
L. M. Henkes et al. / Bioorg. Med. Chem. 20 (2012) 985–995
This vindicates a strong demand for new chemical entities exhibit-
ing optimized HDAC isoform selectivity profiles.6 All HDACi of the
hydroxamic acid type consist structurally of a cap moiety attached
to a mostly aliphatic carbon linker and a zinc binding group (e.g.,
hydroxamic acid moiety) at the opposing side of the linker. In
the attempt to synthesize new HDACi’s, the zinc binding groups
and different capping structures have been extensively varied,
while the spacer was almost exclusively an alkyl chain of different
length.8 Though the importance of the linker component of HDAC-
i’s has been recognized recently9,10, the chemistry of spacers has
just been varied sparsely. This inspired us to substitute the sube-
royl linker of SAHA by a perfluorinated analog and to analyze the
effect on inhibitory activity and selectivity. Compounds with per-
fluorinated alkyl spacer turned out to be inhibitors of HDAC-like
amidohydrolase (HDAH) from Bordetella11 with similar IC50 values
as SAHA. To further explore the effect of hydroxamic acids with
perfluorinated alkyl spacer on human HDACs and to possibly
induce distinct isoform selectivity, a series of analogs with differ-
ent aromatic and heteroaromatic capping groups was synthesized
and tested against a panel of representative recombinant human
classes I, IIa and IIb HDACs and lysates of human K562 cells.
spacer with hydroxamic acid function at the capping group on
enzyme inhibition, biphenyl derivatives 11, 18 and 20 were
synthesized as well. Based on their strong similarities to the potent
HDACi SAHA, these were the first structures we started with. To
further evaluate the sterical influence of the capping group on
the inhibitory potency and selectivity, we synthesized perfluori-
nated analogs with aromatic capping groups of different sizes
( 17, 19, 21, 22, 23). In former studies, aromatic residues with a
maximum of two rings were frequently used, while larger moieties
remained uninspected.8 Heterocycles were already found to mod-
ulate the inhibitory potential in several SAHA analogs.8 Therefore,
the triazole containing compounds 13 and 16 were synthesized. In
contrast to these compounds, the thiazole heterocycle 15 was di-
rectly attached to the perfluorinated spacer. SAHA (25) was used
as a general reference HDACi for all assays. To exclude unspecific
contributions of the perfluorinated spacer to the observed interac-
tion with HDAH and HDACs, compound 24 that lacked the zinc
chelating hydroxamate group was used as a negative control
compound.
2.2. Biochemical and biological characterization
In order to evaluate the inhibitory potential of the newly cre-
ated perfluorinated compounds 4–24, fluorogenic enzyme activity
assays were conducted to obtain IC50 values for different recombi-
nant human HDAC isoforms and the bacterial HDAC homolog
HDAH. The HDAC classes are represented by HDAC1, HDAC8 (class
I), HDAC6 (class IIb) and HDAC7 (class IIa). While HDAC1 and
HDAC8 are still members of the same class, they clearly differ in
their selectivity for substrates and inhibitors.13 The competitive
dual parameter binding assay was used to obtain more information
about the binding site of the representative compound 4. For fur-
ther compound characterization cancer cell line experiments have
2. Results
2.1. Chemistry
The reported standard synthesis of SAHA12 via the formation of
a cyclic anhydride did not work with dodecafluorooctanedioic acid,
probably due to the high electronegativity of the fluorine atoms.
Instead of the anhydride the acid dichloride 2 was synthesized.
Another modification to the SAHA synthesis applied to the inter-
mediate synthesis of the benzylester 3 which served as chemical
starting point for the synthesis of all perfluorinated compounds
(see Scheme 1). One of the chlorides of 2,2,3,3,4,4,5,5,6,6,7,7-dod-
been performed. All compounds of interest (IC50 <10
tested in lysates of K562 cancer cells using the acetylated and tri-
fluoracetylated substrates Boc- -Lys(Ac)-MCA and Boc- -Lys(TFA)-
lM) were also
ecafluorooctanedioyldichloride
2 was substituted by benzyl
L
L
alcohol, leading to benzyl-7-chlorocarbonyl-2,2,3,3,4,4,5,5,6,6,7,
7-dodecafluoroheptanoate 3. For the introduction of several new
cap groups the obtained benzyl 7-chlorocarbonyl-2,2,3,3,4,4,5,
5,6,6,7,7-dodecafluoroheptanoate was reacted with different aro-
matic amines in the presence of a base. The base may be an inor-
ganic or organic base like sodium carbonate, triethylamine or
pyridine. The hydroxamic acid function is introduced by a final
reaction with hydroxylamine in form of its inorganic salt, aqueous
solution in THF or any miscible solvent yielding the desired
hydroxamic acid.
MCA to measure the intrinsic HDAC activity of the K562 cells.
2.2.1. Inhibition of recombinant HDAC enzymes
In a first series, two-step enzyme activity assays with control
compounds 24 and 25 were accomplished using HDAH, HDAC1,
HDAC6, HDAC7 and HDAC8. All Kd-values were summarized in
Table 1. Obtained results confirm SAHA (25) as a suitable reference
inhibitor for HDAH, HDAC1, HDAC6 and HDAC8. Class IIa HDAC7
was only weakly inhibited by SAHA which is in accordance to Brad-
ner et al.10 The dianilide reference compound 24 was only a very
weak inhibitor of HDAH (55 5.5 lM). In accordance to SAHA,
compound 4 showed a significant inhibition of HDAH, HDAC1,
HDAC6 and HDAC7. HDAC8 is lesser favored with an Kd-value of
Perfluorinated hydroxamic acids were divided into four main
groups due to the chemical structure of the capping group. The
Aniline like cap structures 4–10 and 12 represent the largest group
of our set. To analyze the positional effect of the perfluorinated
Scheme 1. Synthesis scheme for hydroxamic acid compounds with perfluorinated spacer. Compound 3 served as chemical starting point for the synthesis of all
perfluorinated hydrocamic acids.

L. M. Henkes et al. / Bioorg. Med. Chem. 20 (2012) 985–995
987
Table 1
Selectivity profile of perfluorinated hydroxamic acids against HDAH and different HDAC isoforms
ID
HDAH
HDAC1
HDAC6
HDAC7
HDAC8
4
5
0.54 0.02
0.27 0.02
29 5.3
14 1.1
11 2.1
11 1.0
52 38
17 0.6
34 5.3
>100
6
7
0.21 0.02
1.3 0.23
12 2.9
18 2.8
23 4.7
19 6.3
13 5.4
35 2.8
36 4.6
9.2 1.7
8
9
0.23 0.01
0.39 0.03
18 1.3
14 8.8
71 35
32 4.1
35 3.8
36 5.4
4.2 1.0
4.1 0.37
10
11
0.26 0.02
8.8 1.0
24 13
3.3 0.52
14 1.0
3.8 1.1
31 13
49 36
14 0.8
57 17
12
13
0.19 0.02
7.0 1.0
16 4.4
1.3 0.22
8.2 2.0
3.0 0.41
30 6.6
13 1.4
>100
>100
14
15
0.6 0.2
2.2 0.2
22 5.8
2.6 0.2
17 4.3
5.4 1.3
75 33
47 11
1.4 0.17
1.5 0.28
16
17
4.5 0.5
66 6.5
47 9.2
12 2.2
>100
0.37 0.03
29 2.6
32 8.3
20 4.1
19 5.9
(continued on next page)

988
L. M. Henkes et al. / Bioorg. Med. Chem. 20 (2012) 985–995
Table 1 (continued)
ID
HDAH
HDAC1
HDAC6
HDAC7
HDAC8
18
0.45 0.05
30 5.7
15 4.7
14 0.9
60 25
19
20
21
0.16 0.02
0.86 0.05
9.1 0.8
12 0.3
49 6.0
65 14
6.8 1.75
9.2 0.9
11 0.8
14 2.9
38 8.3
49 12
5.3 0.47
60 39
>100
22
0.29 0.04
18 2.3
10 1.9
11 1.4
5.3 0.8
23
24
0.74 0.06
27 2.7
14 1.2
5.9 0.73
13 2.2
1.4 0.25
>100
>100
>100
>100
25
0.30 0.02
0.04 0.01
0.30 0.07
36 16
2.9 0.20
The values correspond to Kd-values (in lM) of the corresponding compound tested against one of the denoted purified recombinant HDAC homologues determined by using
fluorogenic enzyme activity assays as described in Section 5.
17 0.6
l
M. While 4 inhibited HDAH with almost the same
meta-Substitution had only a negative impact on inhibition of
HDAC8, whereas the relatively bulky ortho-substitution caused
weaker inhibition of HDAC1 but improved the inhibition of HDAC8.
Apart from a slight decreased potency of 19 against HDAC8 the
naphthalene derivatives 17 and 19 showed only moderate differ-
ences in IC50 values to SAHA analog 4. When proceeding to the en-
larged ring system of anthracene, the difference between the
substitution in position 1 (23) and 2 (21) became more obvious.
Interestingly, the potency of the potentially bulkier 23 remained
almost the same against all HDAC homologs except for HDAC8
when compared with 4. Though being much bulkier, 23 inhibited
HDAC8 10 times stronger than 4. On the other hand the potentially
more elongated compound 21 exhibited pronounced lower po-
tency against all HDAC homologs.
potency than SAHA, HDAC1 was inhibited 725 times and HDAC6
37 times weaker when compared with SAHA. And 4 is 6 times less
potent against HDAC8 than SAHA. In contrast, 4 inhibits class IIa
HDAC7 with about three times higher potency than SAHA indicat-
ing a completely different selectivity profile of both analogs. While
introduction of one p-methyl group (5) caused a slight preference
for HDAC1 and HDAH, substitution with a second methyl residue
in meta position (10) led to a significant inhibition of HDAC7 at a
Kd-value of 3.3 0.52 lM (Table 1). Compounds with para-chlorine
(6) or para-iodine (9) substitutions inhibited HDAH and HDAC1
similarly. However, iodine (9) induces a noticeable inhibition for
HDAC6 (4.2 1.0 lM) and HDAC7 (4.1 0.37 lM) as well. Com-
pound 7 having an acetyl amino group shows a moderate inhibi-
tion of all investigated HDAC homologs except for HDAC8. The
introduction of a para-cyano group (8) resulted in a diminished
inhibition of class II HDACs. The steric conditions at the outer
rim of the active site of the HDAC homologs were further explored
using biphenyl capping groups that were substituted in para- (11),
meta- (18) and ortho- (20) positions of one phenyl ring by the per-
fluorinated linker with terminal hydroxamic acid functional group.
It turned out that para-substitution in general had an unfavorable
impact on the potency of all HDAC homologs except for HDAC7.
Compounds 13 and 16 containing triazol capping groups
showed reduced potency on most HDAC isoforms as compared to
compound 4. Both compounds showed similar decreased potency
against the bacterial HDAH and no or very low inhibition of class
I HDACs 1 and 8 (Fig. 1). On the other hand compound 13 showed
a moderate selectivity for HDAC6 (Kd = 8.2 2.0
lM) whereas com-
pound 16 showed slight selectivity for HDAC7 (Kd = 12 2.2
l
M).
The thiazole containing compound 15 is a strong but rather unse-
lective inhibitor of all investigated HDAC homologs. In addition, it

L. M. Henkes et al. / Bioorg. Med. Chem. 20 (2012) 985–995
989
Figure 1. Selectivity profiles of selected perfluorinated SAHA analogs. Dose–response curves were performed using the enzyme-activity assays as described in Section 5. The
normalized enzyme activity refers to a reaction mixture in the absence of inhibitor compounds and is plotted versus the concentration of the respective compound.
could be shown, that 4 displaced the fluorescent Atto-hydroxamate
ligand from its location of binding within the active site of HDAH in
a dose-dependent manner (Fig. 2) confirming the active site to be
the local target of interaction for the inhibitor. The binding con-
stant between HDAH and 4 was calculated to be 0.4 0.1 lM
according to the method of Riester et al.14
2.2.2. Inhibition of cellular HDACs
Selected by potency, compounds 4, 7, 9, 10, 12, 15, 19, 20, 23
were subjected to a two-step-assay in the presence of diluted cell
lysate, instead of purified recombinant human HDACs. The cell
lysate was obtained under natural conditions from human K562
cells. The activity of all HDAC isoforms was addressed by two sub-
strates which are known to be processed by either class I HDACs
1,2,3 and HDAC 6 (Boc-
L
-Lys(acetyl)-MCA) or class IIa HDACs
4,5,7,9 and HDAC 8 (Boc-
L
-Lys(trifluoroacetyl)-MCA).15
The resulting IC50 values were summarized in Table 2. Only
compound 15 showed significant but rather unselective inhibition
of the cellular HDACs using both substrates. The corresponding IC50
Figure 2. Binding isotherm of compound 4 to HDAH. The binding of compound 4 to
2.5 M HDAH was assessed at 21 °C using the dual parameter competition assay
under standard assay condition employing 50 nM of Atto-ligand probe and denoted
concentrations of compound 4. The change of fluorescence lifetime of the
fluorescent probe upon its displacement from HDAH indicated binding of com-
pound 4. The straight line represents a data fit to a competitive binding model
l
values were 30 3.2
lM for Boc-L-Lys(acetyl)-MCA and 14 1.7
l
M for Boc- -Lys(trifluoroacetyl)-MCA, respectively.
L
yielding a binding constant of 0.4 lM.
2.3. Molecular docking and simulation study
Blind docking of reference compound SAHA and its perfluori-
nated derivative 4 into the 3D-structure of a homology model of
HDAC1 and X-ray structures of HDAH (1ZZ1) and HDAC8 (1T67)
was performed to look for potential binding sites and to obtain a
suitable starting point for subsequent molecular dynamics (MD)
simulations. The ligand structures of more than 100 docking poses
were clustered according to RMSD values for each protein–ligand
complex. The energetically most favorable clusters of both, SAHA

990
L. M. Henkes et al. / Bioorg. Med. Chem. 20 (2012) 985–995
Table 2
Inhibition of native HDACs from K562 cell lysate
Compounds
Boc-Lys-(acetyl)-MCA
Boc-Lys-(trifluoroacetyl)-MCA
4
7
>50
>50
49 2.9
>50
9
>50
>50
10
12
15
19
20
23
>50
>50
30 3.2
>50
>50
>50
42 3.4
14 1.7
45 3.8
>50
>50
>50
The values correspond to IC50 values (in lM) of the corresponding compound tested
in the presence of K562 cell lysate using two different fluorogenic substrates
(Boc-Lys-(acetyl)-MCA and Boc-Lys-(trifluoroacetyl)-MCA).
and 4, were always located in the active sites of HDAH and human
HDACs 1 and 8 (see Supplementary Figs. S1–S3). For both com-
plexes, HDAH with SAHA and 4, the corresponding energetically
most favorable poses were taken as starting points for energy min-
imization and cascading MD simulations at increasing tempera-
tures according to the protocol described in Section 5. The
potential energies of both complexes showed no skewness over
the final 2 ns MD run but rather scattered around a constant value,
indicating an equilibrated system (see Supplementary Fig. S4). The
structurally aligned protein structures achieved equilibrium after
about 500 ps when no significant changes of the root mean square
distance (RMSD) could be observed. In contrast, the RMSD values
of the small ligands fluctuated much more reflecting the flexibility
of the free rotatable phenyl capping group and to a lesser extent
the spacer (see Supplementary Fig. S5). The final complex struc-
tures with both ligands were matched by minimizing the RMSD
between the corresponding proteins. Both ligands were shown to
form bidentate complexes with the catalytical Zn2+ and to have
similar orientations within the active site (Fig. 3).
3. Discussion
SAHA is a rather unselective inhibitor for all class I and class IIb
HDACs10 and an FDA approved cancer drug. Actually, SAHA is also
investigated in numerous clinical trials16 to extend its clinical indi-
cation from cutaneous T cell lymphoma to other types of cancer.
Based on its chemical structure a new scaffold with perfluorated
alkyl chain was created (Scheme 1). The linker region was sup-
posed to change its electronic charge distribution and hydrophobic
character, due to the substitution of the 12 hydrogen atoms by
highly electronegative flourine atoms. This became already notice-
able during synthesis of the activated perfluorinated spacer inter-
mediate which could not be obtained as anhydride like in the
standard synthesis of SAHA12 but rather as an acid dichloride.
However, compound 4 as the closest analog to SAHA showed
noticeable inhibition of HDAH and also of all tested recombinant
HDACs even though its selectivity profile was clearly distinct from
SAHA. While the potency of 4 and SAHA against HDAH was compa-
rable, compound 4 turned out to inhibit human classes I and IIb
HDACs 6–360 times weaker whereas 4 was about 3 times more po-
tent against class IIa member HDAC7. This was the first hint that
the perfluorated scaffold could contribute to class II selectivity.
The negative control 24 lacking the hydroxamate moiety achieved
no significant inhibition of any HDAC homolog indicating no obvi-
ous unspecific contribution of the newly introduced perfluorinated
spacer. Obtained results from 4 and 24 necessitate the presence of
the Zn2+ chelating hydroxamate group for efficient inhibition of
HDACs. The graduated IC50 values of the synthesized compounds
5–23 with respect to different HDAC isoform from classes I, IIa
Figure 3. MD-simulations of protein–ligand complexes consisting of HDAH and
SAHA or its perfluorinated analog (4). The inhibitor molecules were docked into the
3D-structure of HDAH (PDB 1ZZ1 chain C) and subsequently subjected to MD-
simulations. The final complex structures were matched by minimizing the root
mean square deviation between the proteins. (A) Final pose of MD-simulation:
Perfluorinated SAHA formed a bidental complex with the Zn2+ ion (gray bead) at the
bottom of the active site of HDAH. (B) Overlay of the active site of HDAH in complex
with SAHA (blue) and its perfluorinated analog 4 (red).
and IIb and HDAH proved the potential of SAHA analogs with a per-
fluorinated substructure (Table 1).
Moreover, compounds differing in their cap groups, developed
distinct selectivity profiles with varying potencies. In general, com-
pounds 5–10 and 12 with smaller substituents exhibited similar
IC50 values for HDAH, and class I HDACs 1 and 8. The iodine and
ethyl substituted analogs 9 and 12, respectively, showed a moder-
ate preference for class II HDACs 6 and 7 confirming the potential
to create class II selectivity using the perfluoralkyl-hydroxamate
scaffold. Three of our compounds (5, 10 and 12) directly corre-
sponded to compounds 4, 12 and 6 in Oger’s study17 of SAHA
analogs differing only in the perfluorination of the spacer. Interest-
ingly, the moderate class IIa/IIb preference of our compounds 10
and 12 was not observed for the corresponding unfluorinated
counterparts. Moreover, the SAHA analogs with plain alkyl chains
exhibited a distinct selectivity for HDAC1 and HDAC6 but not for

L. M. Henkes et al. / Bioorg. Med. Chem. 20 (2012) 985–995
991
HDAC7. This direct comparison between perfluorated and unflu-
orated SAHA analogs suggests that the perfluorinated linker itself
contributes to the induction of selectivity for HDAC class II, partic-
ularly class IIa.
Inspired by these results and assuming an inhibition binding
mode similar to that of SAHA, we tried to generate even stronger
differentiating selectivity profiles by synthesizing perfluorinated
SAHA analogs with bulkier substitutions at the aromatic ring.
Trying biphenyl capping groups (11, 18, 20) differently attached
to the perfluorinated hydroxamate scaffold the para-substitution
interfered with the activity of all HDACs except for HDAC7 and
ortho-substitution increased the potency of 20 against HDAC8 to
sites of HDAH, HDAC1 and HDAC8. Subsequent MD simulations
demonstrated similar orientations of SAHA and 4 within the active
site of HDAH. However, the ligand structures showed pronounced
fluctuations indicating the necessity of prolonged MD simulations
to improve the understanding of the molecular base of ligand–en-
zyme interaction. In the competitive binding assay compound 4,
the perfluorinated analog of SAHA, acted as a competitive ligand
displacing the fluorescent probe Atto-hydroxamic acid, which has
been shown to bind to the Zn2+ containing active site of HDAH14
,
from its binding site (Fig. 2). This finding confirmed the blind dock-
ing and MD simulation results with HDAH and provided strong evi-
dence that 4 and presumably the other perfluorinated analogs of
SAHA, as well, bind to the active site of HDAH and probably also
an Kd-value of 5.3 0.47 lM. Using differently attached naphtha-
lene capping groups (17, 19) also showed no significant improve-
ment of compound selectivity profiles. The bulkier anthracene
capping group attached at position 2 (21) caused a drop in potency
against HDAH and all human HDACs. When anthracene is attached
to the perfluorinated hydroxamate scaffold via position 1 (23), the
IC50 values of HDAH, HDAC1, HDAC6 and HDAC7 become similar to
the closest SAHA analog compound 4. Moreover, 23 was an about
10 times more potent inhibitor of HDAC8 than 4 indicating that
the anthracene capping group undergoes additional presumably
hydrophobic interaction with the surface of HDAC8. Combining
the results of the compounds with biphenyl and anthracene cap-
ping groups, more elongated structures are supposed to lead to
unfavorable HDAC inhibitors, in general. On the other hand, more
bulkier cap group substitutions seemed to favor the inhibition of
HDAC8 and in general not to decrease the inhibitory potency
against the other HDAC homologs. The introduction of capping
groups containing triazol rings leads to compounds 13 and 16
and reduced potency against HDAH and class I HDACs. Both com-
pounds appeared to be moderately selective for class II HDACs.
Since 13 showed a moderate preference for HDAC6 whereas 16
was slightly selective for HDAC7, systematic variations of triazol
based capping groups may enable to tune the selectivity profile
with respect to class IIa or IIb. Compound 15 is the only compound
in our series of novel HDAC inhibitors where the phenyl ring of
perfluorinated SAHA analogs is replaced by a thiazole heterocycle
that is substituted by another phenyl ring. Compound 15 turned
out to be an unselective and comparable strong inhibitor of all hu-
man HDACs and HDAH even though its potency against HDAC1 and
HDAC6 is much lower than that of SAHA. On the other hand 15 was
found to be about 25 times more potent against HDAC7 and twice
as potent against HDAC8 than SAHA. Consistent with its allover
inhibitory effect on all recombinant HDACs, 15 was able to inhibit
the processing of two different substrates with complementary
isoform selectivity by native HDAC multi-enzyme complexes in
K562 cellular lysate. The IC50 values of, for example 15 or 12 in
the lysate of K562 cells were significantly higher compared with
those of the biochemical assay using recombinant HDACs. In fact,
only 15 was found to be a clear but rather weak inhibitor of human
HDACs in the cell lysate. This might be explained by an increased
disposition for unspecific binding to the surface of hydrophobic
proteins. This hypothesis is backed by our finding that addition
of BSA to the enzyme activity assay using purified recombinant
HDACs clearly reduced the potency of perfluorinated SAHA analogs
in vitro (unpublished results).
of human HDACs. The corresponding binding constant of 0.4
lM
was in good agreement with the Kd-value (0.5 M, see Table 1)
l
calculated from the IC50-value of the enzyme activity assay.
4. Conclusion
In summary, we proved the potential of compounds with per-
fluorinated alkyl spacer as HDAC inhibitors. Binding displacement
experiments and combined blind docking and MD simulations sug-
gest the perfluorinated analogs of SAHA to bind to the active site of
HDAH as well as human HDACs. The replacement of the ordinary
alkyl spacer in SAHA and its analogs by perfluorinated alkyl chains
created completely different selectivity profiles against different
HDAC isoforms. In contrast to the alkyl spacer of SAHA and deriv-
atives, the perfluorinated alkyl spacer seems to contribute to or
facilitate the induction of selectivity for class II, particularly class
IIa, HDACs even though the overall potency of the perfluorinated
SAHA analogs in this study against human HDACs remained still
rather moderate. However, when designing novel selective class
II inhibitors attention should not exclusively be paid to the capping
group and to the distance between capping and zinc binding group
but also to the chemistry of the spacer itself, particularly consisting
of partially or fully fluorinated linker structures.
5. Experimental section
TLC was performed on TLC aluminium sheets—Silica Gel 60
F254 (Merck). Column chromatography was performed on silica
gel 40–63 mesh, (Merck). 1H NMR-spectra were measured on a
Bruker Avance (300.13 MHz) and the 19F NMR-spectra were mea-
sured on a Bruker DRX-400 (376.33 MHz) in CDCl3 or DMSO-d6
and referenced to TMS. 19F NMR-spectra were referenced to hexa-
fluorobenzene as external standard. Mass spectra were measured
on a Finnigan LCQDECA (Thermoquest) using APCI (negative mode).
Chemicals and solvents were purchased from Sigma–Aldrich, VWR,
Alfa Aesar and Fisher Scientific.
5.1. General synthesis for 2,2,3,3,4,4,5,5,6,6,7,7-dodecafluo-
rooctanedioyldichloride
2,2,3,3,4,4,5,5,6,6,7,7-Dodecafluorooctanedioic acid (25.00 g,
64.09 mmol) was poured into a flame dried three necked flask
cooled under a nitrogen atmosphere. Thionylchloride (28.0 ml,
45.67 g, 383.9 mmol) was added via a dropping funnel and DMF
(8 drops) was added carefully to the colorless suspension. The reac-
tion mixture was refluxed for 4 h. After completion of the reaction
remaining thionylchloride was distilled off. The product was ob-
tained as colorless oil by subsequent distillation of the residue un-
der reduced pressure (bp 93–96 °C/100 mbar). Yield: 23.46 g
(54.95 mmol); 86%; C8Cl2F12O2; M = 426.97 g/mol; 19F NMR
(CDCl3): d (ppm) = À117.6–(À117.8) [m, 4F]; À125.8–(À126.0)
[m, 8F].
The inhibition of HDAH and human HDACs by perfluorinated
analogs of SAHA has been proven but the exact location of the
binding site was not clear a priori. In principle, inhibitors could
bind to the active site or to other allosteric sites on the surface of
an enzyme. The question of the actual binding site was examined
by blind docking, molecular dynamics (MD) simulations and
experimentally by a competitive binding assay using a fluorescent
Atto-hydroxamate ligand probe. Blind docking revealed a definite
preference of both, SAHA and its analog 4, for binding to the active

992
L. M. Henkes et al. / Bioorg. Med. Chem. 20 (2012) 985–995
5.2. General synthesis of benzyl-7-chlorocarbonyl-
2,2,3,3,4,4,5,5,6,6,7,7-dodecafluoroheptanoate
Within 30 min a solution of triethylamine (0.17 ml, 0.12 g,
1.19 mmol) and the particular aromatic amine or heteroaromatic
amine (1.00 mmol), respectively in the required amount of THF
(10–20 ml) was added via a dropping funnel, maintaining the reac-
tion temperature below 5 °C. The reaction mixture was stirred for
2 h. After additional stirring for 3 h at room temperature 1.0 M
hydroxylamine (2.00 ml, 2.00 mmol) in THF was added via syringe.
The reaction mixture was stirred at room temperature until TLC
indicated complete consumption of the intermediate. The solvent
was evaporated and the product was isolated by column
chromatography.
5.2.1. The general synthesis followed two different standard
procedures, described below
5.2.1.1. Procedure 1. 2,2,3,3,4,4,5,5,6,6,7,7-Dodecafluoro-
octanedioyldichloride (23.46 g, 54.95 mmol) was poured into a
flame dried three necked flask cooled under a nitrogen atmo-
sphere. The starting material was cooled in an ice bath, before ben-
zyl alcohol (3.40 ml, 3.55 g, 32.83 mmol) was added via a dropping
funnel under vigorous stirring over a period of 1 h. The ice bath
was removed and the reaction mixture was heated at 100 °C for
one additional hour. After cooling to room temperature excessive
starting material was collected by distillation under reduced pres-
sure (bp 93–96 °C/100 mbar). High vacuum distillation of the resi-
due provided the product as colorless oil (bp 90–92 °C/0.1 mbar).
Yield: 7.70 g (15.44 mmol); 47%
5.3.3. 2,2,3,3,4,4,5,5,6,6,7,7-Dodecafluorooctanedioic acid
hydroxyamide phenyl-amide (4)
Using aniline (0.10 ml; 0.10 g; 1.07 mmol) as starting material
the product was obtained as a colorless solid according to standard
procedure 2.
Yield: 0.07 g (0.15 mmol); 14%
C14H8F12N2O3; M = 480.21
g/mol; 1H NMR (DMSO-d6): d (ppm) = 7.28–7.38 (m, 1H); 7.46–
7.53 (m, 2H); 7.71–7.81 (m, 2H); 9.97 (s, br, 1H); 11.36 (s, br,
1H); 12.41 (s, br, 1H); 19F NMR (DMSO-d6): d (ppm) = À117.5–
(À117.8) [m, 2F]; À118.1–(À118.5) [m, 2F]; À120.7–(À121.3) [m,
4F]; À121.5–(121.7) [m, 2F]; À121.8–(À122.0) [m, 2F]; MS (Àp
APCI): m/z = 479 (100%) [MÀ1]À.
5.2.1.2. Procedure 2. 2,2,3,3,4,4,5,5,6,6,7,7-Dodecafluoro-
octanedioyldichloride (18.16 g, 42.53 mmol) was dissolved in
50 ml dioxan and poured into a flame dried three necked flask
cooled under a nitrogen atmosphere. The colorless solution was
heated in an oil bath at 50 °C. A solution of benzyl alcohol
(3.10 ml, 3.24 g, 29.96 mmol) in 25 ml dioxan was added drop wise
over a period of 90 min. The reaction mixture was stirred at 100 °C
for additional 20 h. After cooling to room temperature the solvent
was removed by distillation in vacuo. High vacuum distillation of
the residue provided the product as colorless oil (bp 90–92 °
C/0.1 mbar).
Yield: 6.83 g (13.70 mmol); 46%; C15H7ClF12O3; M = 498.65 g/
mol; 1H NMR (CDCl3): d (ppm) = 5.27 [s, 2H]; 7.24–7.41 [m, 5H];
19F NMR (CDCl3): d (ppm) = À113.9–(À114.1) [m, 2F]; À119.4–
(À119.6) [m, 2F]; À122.1–(À122.5) [m,4F]; À122.6–(À122.9) [m,
2F]; À123.5–(À123.8) [m, 2F].
5.3.4. 2,2,3,3,4,4,5,5,6,6,7,7-Dodecafluorooctanedioic acid
hydroxyamide 4-methyl-phenylamide (5)
Using 4-methylaniline (0.11 ml; 0.11 g; 1.03 mmol) as starting
material the product was obtained as a yellowish solid according
to standard procedure 2.
Yield: 0.09 g (0.18 mmol); 17%; C15H10F12N2O3; M = 494.23
g/mol; 1H NMR (DMSO-d6): d (ppm) = 2.34 (s, 3H); 7.21–7.29 (m,
2H); 7.52–7.60 (m, 2H); 9.97 (s, br, 1H); 11.19 (s, br, 1H); 12.36
(s, br, 1H); 19F NMR (DMSO-d6): d (ppm) = À118.1–(À118.3) [m,
2F]; À119.4–(À119.7) [m, 2F]; À121.5–(À121.9) [m, 4F]; À122.2–
(À122.7) [m, 4F]; MS (Àp APCI): m/z = 493 (100%) [MÀ1]À.
5.3. Cap group synthesis
5.3.1. Introduction of new cap groups was accomplished by the
following standard procedures
5.3.5. 2,2,3,3,4,4,5,5,6,6,7,7-Dodecafluorooctanedioic acid (4-
chlorophenyl)amide hydroxyamide (6)
Using 4-chloroaniline (0.13 g; 1.02 mmol) as starting material
the product was obtained as a colorless solid according to standard
procedure 2.
Yield: 0.07 g (0.14 mmol); 14%; C14H7ClF12N2O3; M = 514.65
g/mol; 1H NMR (DMSO-d6): d (ppm) = 7.48–7.53 (m, 2H); 7.69–
7.78 (m, 2H); 9.93 (s, br, 1H); 11.41 (s, br, 1H); 12.37 (s, br, 1H);
19F NMR (DMSO-d6): d (ppm) = À119.5–(À119.7) [m,2F]; À119.8–
(À120.2) [m, 2F]; À121.7–(À122.0) [m, 4F]; À122.3–(À123.1) [m,
4F]; MS (Àp APCI): m/z = 515 (33%) [MÀ1]À; 513 (100%) [MÀ1]À.
5.3.1.1. Standard procedure 1. 5.3.1.1.1. First step. Benzyl-
7-chlorocarbonyl-2,2,3,3,4,4,5,5,6,6,7,7-dodecafluoroheptanoate
(0.50 g, 1.00 mmol) was dissolved in 10 ml THF and poured into a
flame dried three necked flask cooled under a nitrogen atmosphere.
The colorless solution was cooled in an ice bath. Within 30 min a
solution of triethylamine (0.17 ml, 0.12 g, 1.19 mmol) and the
particular aromatic amine or heteroaromatic amine (1.00 mmol),
respectively in the required amount of THF (10–20 ml) was added
via a dropping funnel, maintaining the reaction temperature below
5 °C. The reaction mixture was stirred for 2 h, before the ice bath
was removed. After additional stirring for 3 h the solvent was
evaporated and the intermediate was purified by column
chromatography.
5.3.6. 2,2,3,3,4,4,5,5,6,6,7,7-Dodecafluorooctanedioic acid (4-
acetylaminophenyl)-amide hydroxyamide (7)
Using 4-acetylaminoaniline (0.15 g; 1.00 mmol) as starting
material the product was obtained as a yellowish solid according
to standard procedure 2.
Yield: 0.07 g (0.11 mmol); 11%; C16H11F12N3O4; M = 537.26
g/mol; 1H NMR (DMSO-d6): d (ppm) = 2.18 (s, 3H); 7.60–7.82 (m,
4H); 10.03 (s, br, 1H); 10.13 (s, 1H); 11.27 (s, br, 1H); 12.44 (s,
br, 1H); 19F NMR (DMSO-d6): d (ppm) = À118.7–(À119.1) [m, 2F];
À 119.5–(À119.9) [m, 2F]; 122.2–(À123.1) [m, 8F]; MS (Àp APCI):
m/z = 536 (100%) [MÀ1]À.
5.3.1.1.2. Second step. A solution of the particular benzyl-
7-arylcarbamoyl-2,2,3,3,4,4,5,5,6,6,7,7-dodecafluoroheptanoate or
benzyl-7-heteroarylcarbamoyl-2,2,3,3,4,4,5,5,6,6,7,7-dodecafluoro-
heptanoate, respectively in 10 ml abs THF (cres = 0.043 mmol/ml)
was treated via syringe with the required amount of 1.0-M hydrox-
ylamine (2.0 equiv) in THF. The reaction mixture was stirred at
room temperature for 24 h. After evaporation of the solvent the
product was isolated by column chromatography.
5.3.2. Standard procedure 2
5.3.7. 2,2,3,3,4,4,5,5,6,6,7,7-Dodecafluorooctanedioic acid (4-
cyanophenyl)amide hydroxyamide (8)
Using 4-aminobenzonitrile (0.12 g; 1.02 mmol) as starting
material the intermediate Benzyl 7-(4-cyanophenylcarbamoyl)-
Benzyl-7-chlorocarbonyl-2,2,3,3,4,4,5,5,6,6,7,7-dodecafluoroh-
eptanoate (0.50 g, 1.00 mmol) was dissolved in 10 ml THF and
poured into a flame dried two necked flask cooled under a nitrogen
atmosphere. The colorless solution was cooled in an ice bath.

L. M. Henkes et al. / Bioorg. Med. Chem. 20 (2012) 985–995
993
2,2,3,3,4,4,5,5,6,6,7,7-dodecafluoroheptanoate was obtained as a
colorless solid according to standard procedure 1 step 1.
Yield: 0.15 g (0.26 mmol); 26%; C22H12F12N2O3; M = 580.07
g/mol; 1H NMR (DMSO-d6): d (ppm) = 5.58 (s, 2H); 7.45–7.55 (m,
5H); 7.95–8.03 (m, 4H); 11.72 (s, br, 1H); 19F NMR (DMSO-d6): d
(ppm) = À118.1–(À118.4) [m, 4F]; À121.3–(À121.7) [m, 4F];
À122.0–(À122.2) [m, 2F]; 122.4–(À122.7) [m, 2F];MS (Àp APCI):
m/z = 579 (100%) [MÀ1]À.
(ppm) = À116.9–(À117.2) [m, 2F]; À117.6–(À117.9) [m, 2F];
À120.3–(À120.6) [m, 4F]; À121.0–(À121.3) [m, 2F]; À121.3–
(À121.6) [m, 2F]; MS (Àp APCI): m/z = 507 (100%) [MÀ1]À.
5.3.12. 2,2,3,3,4,4,5,5,6,6,7,7-Dodecafluorooctanedioic acid
hydroxyamide (4-[1-phenyl-1H-(1,2,3)triazol-4-
yl]phenyl)amide (13)
Using 4-(4-aminophenyl)-1-phenyl-1H-[1,2,3]-triazole (0.24 g;
1.02 mmol) as starting material the product was obtained as a
colorless solid according to standard procedure 2.
Using benzyl-7-(4-cyanophenylcarbamoyl)-2,2,3,3,4,4,5,5,6,6,
7,7-dodecafluoroheptanoate (0.13 g; 0.22 mmol) as starting
material the product was obtained as a colorless solid according
to standard procedure 1 step 2.
Yield: 0.23 g (0.37 mmol); 36%; C22H13F12N5O3; M = 623.35
g/mol; 1H NMR (DMSO-d6): d (ppm) = 7.39–7.56 (m, 2H); 7.59–
7.71 (m, 2H); 7.78–7.88 (m, 2H); 7.92–8.02 (m, 5H); 9.32 (s, 1H);
Yield: 0.04 g (0.08 mmol); 36%;
C15H7F12N3O3; M = 505.22
g/mol; 1H NMR (DMSO-d6): d (ppm) = 7.96–8.01 (m, 4H); 10.00 (s,
br, 1H); 11.73 (s, br, 1H); 12.41 (s, br, 1H); 19F NMR (DMSO-d6): d
(ppm) = À118.5–(À118.9) [m, 2F]; À119.2–(À119.6) [m, 2F];
À121.5–(À122.4) [m, 8F]; MS (Àp APCI): m/z = 504 (100%) [MÀ1]À.
11.38 (s, br, 1H); 19F NMR (DMSO-d6):
d
(ppm) = À113.9–
(À114.3) [m, 2F]; À117.2–(À117.6) [m, 2F]; À120.5–(À121.5) [m,
8F]; MS (Àp APCI): m/z = 622 (4%) [MÀ1]À; 607 (100%) [MÀ15]À;
563 (100%) [MÀ59]À.
5.3.8. 2,2,3,3,4,4,5,5,6,6,7,7-Dodecafluorooctanedioic acid
hydroxyamide (4-iodo-phenyl)amide (9)
5.3.12.1. 4-(4-Aminophenyl)-1-phenyl-1H-[1,2,3]-triazole in
Section 5.3.12. was prepared by this way.
A 100 mL 2 necked
Using 4-iodoaniline (0.22 g; 1.00 mmol) as starting material the
product was obtained as a colorless solid according to standard
procedure 2.
flask was charged with a mixture of 25 mL t-butanol and 25 mL
water. Then were added 2.40 g (20.00 mmol) phenylazide and
2.36 g (20.00 mmol; 1 equiv) 4-aminophenylethyne. Afterward
were given 2 mL of a 1 m sodium ascorbate solution (2.00 mmol;
0.1 equiv) and 50 mg of copper-II-sulfate pentahydrate
(0.20 mmol; 0.01 equiv). The dark green suspension was then
heated at 60 °C for 17 h. 50 mL of water were added to the reaction
mixture after cooling down to room temperature. The mixture was
then filtered and washed with water, giving a beige yellow solid,
3.50 g (14.75 mmol); 74%; C14H13N4; M = 237.28 g/mol; 1H NMR
(DMSO): d (ppm) = 5.42 (s, br, 2H); 6.76 (d, 2H); 7.41 (m, 1H);
7.52 (m, 4H); 7.95 (d, 2H); 9.05 (s, 1H); 13C NMR (DMSO): d
(ppm) = 113.94; 117.11; 117.73; 119.74; 126.40; 128.35; 129.83;
136.78; 148.33; 148.95; RF (diethylether): 0.54.
Yield: 0.04 g (0.08 mmol); 8%;
C14H7F12IN2O3; M = 606.10
g/mol; 1H NMR (DMSO-d6): d (ppm) = 7.52–7.71 (m, 2H); 7.81–
7.89 (m, 2H); 10.01 (s, br, 1H); 11.41 (s, br, 1H); 12.43 (s, br,
1H); 19F NMR (DMSO-d6): d (ppm) = À118.5–(À118.8) [m, 2F];
À119.3–(À119.7) [m, 2F]; À121.4–(À121.9) [m, 4F]; À122.0–
(À122.5) [m, 4F]; MS (Àp APCI): m/z = 605 (100%) [MÀ1]À.
5.3.9. 2,2,3,3,4,4,5,5,6,6,7,7-Dodecafluorooctanedioic acid (3,4-
dimethylphenyl)-amide hydroxyamide (10)
Using 3,4-dimethylaniline (0.17 ml; 0.12 g; 0.99 mmol) as start-
ing material the product was obtained as a colorless solid accord-
ing to standard procedure 2.
Yield: 0.04 g (0.08 mmol); 8%;
C
16H12F12N2O3; M = 508.26
5.3.13. 2,2,3,3,4,4,5,5,6,6,7,7-Dodecafluorooctanedioic acid
hydroxyamide pyren-1-ylamide (14)
Using 1-aminopyren (0.22 g; 1.01 mmol) as starting material
the product was obtained as a brownish solid according to stan-
dard procedure 2.
Yield: 0.14 g (0.23 mmol); 23%; C24H12F12N2O3; M = 604.34
g/mol; 1H NMR (DMSO-d6): d (ppm) = 8.03–8.44 (m, 9H); 9.99 (s,
br, 1H); 11.94 (s, br, 1H); 12.43 (s, br, 1H); 19F NMR (DMSO-d6):
d (ppm) = À117.2–(À117.5) [m, 2F]; À118.0–(À118.3) [m, 2F];
À120.5–(À121.1) [m, 4F]; À121.3–(À121.6) [m, 2F]; À121.8–
(À122.1) [m, 2F]; MS (Àp APCI): m/z = 603 (100%) [MÀ1]À.
g/mol; 1H NMR (DMSO-d6): d (ppm) = 2.19–2.28 (m, 6H); 7.11–
7.20 (m, 1H); 7.33–7.40 (m, 1H); 7.42–7.48 (m, 1H); 9.92 (s, br,
1H); 11.11 (s, br, 1H); 12.35 (s, br, 1H); 19F NMR (DMSO-d6): d
(ppm) = À118.2–(À118.5) [m, 2F]; À119.1–(À119.5) [m, 2F];
À121.5–(À121.8) [m, 4F]; À122.0–(122.5) [m, 4F]; MS (Àp APCI):
m/z = 507 (100%) [MÀ1]À.
5.3.10. 2,2,3,3,4,4,5,5,6,6,7,7-Dodecafluorooctanedioic acid
biphenyl-4-ylamide hydroxyamide (11)
Using biphenyl-4-ylamine (0.17 g; 1.00 mmol) as starting mate-
rial the product was obtained as a colorless solid according to stan-
dard procedure 2.
Yield: 0.08 g (0.14 mmol); 14%; C20H12F12N2O3; M = 556.30
g/mol; 1H NMR (DMSO-d6): d (ppm) = 7.31–7.39 (m, 1H); 7.39–
7.53 (m, 2H); 7.67–7.82 (m, 6H); 9.97 (s, br, 1H); 11.36 (s, br,
1H); 12.36 (s, br, 1H); 19F NMR (DMSO-d6): d (ppm) = 116.9–
(À117.1) [m, 2F]; À117.7–(À118.0); [m, 2F]; À120.3–(À120.6)
[m, 4F]; À121.0–(À121.2) [m, 2F]; À121.3–(À121.5) [m, 2F]; MS
(Àp APCI): m/z = 555 (100%) [MÀ1]À.
5.3.14. 2,2,3,3,4,4,5,5,6,6,7,7-Dodecafluorooctanedioic acid
hydroxyamide (4-phenylthiazol-2-yl)amide (15)
Using 2-amino-4-phenylthiazole (0.18 g; 1.02 mmol) as starting
material the product was obtained as a brownish solid according to
standard procedure 2.
Yield: 0.07 g (0.12 mmol); 12%; C17H9F12N3O3S; M = 563.32
g/mol; 1H NMR (DMSO-d6): d (ppm) = 7.37–7.55 (m, 4H); 7.72 (s,
1H); 7.85–7.92 (m, 1H); 9.94 (s, br, 1H); 12.33 (s, br, 1H); 14.45
(s, br, 1H); 19F NMR (DMSO-d6): d (ppm) = À116.3–(À116.8) [m,
2F]; À118.2–(À118.5) [m, 2F]; À120.7–(À121.3) [m, 4F]; À121.3–
(À121.6) [m, 2F]; À121.8–(À122.1) [m, 2F]; MS (Àp APCI):
m/z = 562 (100%) [MÀ1]À.
5.3.11. 2,2,3,3,4,4,5,5,6,6,7,7-Dodecafluorooctanedioic acid
(3-ethylphenyl)amide hydroxyamide (12)
Using 3-ethylaniline (0.12 ml; 0.12 g; 1.00 mmol) as starting
material the product was obtained as a colorless solid according
to standard procedure 2.
5.3.15. 2,2,3,3,4,4,5,5,6,6,7,7-Dodecafluorooctanedioic acid
hydroxyamide [4-(1-phenyl-1H-[1,2,3]triazol-5-
yl)phenyl]amide (16)
Using 5-(4-aminophenyl)-1-phenyl-1H-[1,2,3]-triazole (0.24 g;
1.02 mmol) as starting material the product was obtained as a
brownish solid according to standard procedure 2.
Yield: 0.13 g (0.26 mmol); 26%; C16H12F12N2O3; M = 508.26
g/mol; 1H NMR (DMSO-d6):
d
(ppm) = 1.17–1.21 (t, 3H,
3J = 7.0 Hz); 2.57–2.68 (q, 2H, 3J = 7.0 Hz); 7.07–7.13 (m, 1H);
7.29–7.37 (m, 1H); 7.47–7.55 (m, 1H); 9.95 (s, br, 1H); 11.19 (s,
br, 1H); 12.35 (s, br, 1H); 19F NMR (DMSO-d6):
d

994
L. M. Henkes et al. / Bioorg. Med. Chem. 20 (2012) 985–995
Yield: 0.18 g (0.29 mmol); 28%; C22H13F12N5O3; M = 623.35
À120.2–(À120.7) [m, 4F]; 121.0–(À121.3) [m, 2F]; À121.4–
(À121.7) [m, 2F]; MS (Àp ESI): m/z = 529 (100%) [MÀ1]À.
g/mol; 1H NMR (DMSO-d6): d (ppm) = 7.32–7.38 (m, 2H); 7.41–
7.49 (m, 2H); 7.54–7.61 (m, 3H); 7.68–7.75 (m, 2H); 8.17 (s, 1H);
9.93 (s, br, 1H); 11.38 (s, br, 1H); 12.38 (s, br, 1H); 19F NMR
(DMSO-d6): d (ppm) = À114.6–(À114.9) [m, 2F]; À117.7–(À118.0)
[m, 2F]; À120.8–-121.9 [m, 8F]; MS (Àp APCI): m/z = 622 (9%)
[MÀ1]À; 607 (72%) [MÀ15]À; 563 (100%) [MÀ59]À.
5.3.19. 2,2,3,3,4,4,5,5,6,6,7,7-Dodecafluorooctanedioic acid
biphenyl-2-ylamide hydroxyamide (20)
Using biphenyl-2-ylamine (0.17 g; 1.00 mmol) as starting mate-
rial the product was obtained as a colorless solid according to stan-
dard procedure 2.
5.3.15.1. The used 5-(4-aminophenyl)-1-phenyl-1H-[1,2,3]-tria-
zole in Section 5.3.15 was prepared according to the following
Yield: 0.07 g (0.13 mmol); 13%; C20H12F12N2O3; M = 556.30
g/mol; 1H NMR (DMSO-d6): d (ppm) = 7.31–7.53 (m, 9H); 9.93 (s,
br, 1H); 11.05 (s, br, 1H); 12.35 (s, br, 1H); 19F NMR (DMSO-d6):
d (ppm) = À117.5–(À117.8) [m, 2F]; À118.2–(À118.5) [m, 2F];
À120.8–(À121.3) [m, 4F]; À121.6–(À121.9) [m, 2F]; À121.9–
(À122.2) [m, 2F]; MS (Àp ESI): m/z = 555 (100%) [MÀ1]À.
procedure.
A 100 mL 2 necked flask was charged with 50 mL
of toluene and 2.36 g (20.00 mmol; 1 equiv) 4-aminophenylethyne.
2.40 g (20.00 mmol) phenylazide were then added and the black
solution was heated to reflux for 43 h. The toluene was removed
at reduced pressure and the residue was resolved in 2 N hydrochlo-
ric acid. Unsoluble material was separated and the liquid phase
was washed three times with ethylacetate afterward. The aqueous
phase was then heated with some charcoal, filtrated and neutral-
ized with concd sodium hydroxide solution. The black material
which deposited from the solution was filtrated, dried and crystal-
lized from a mixture of water/ethanol. Brown crystals were then
separated and dried.
5.3.20. 2,2,3,3,4,4,5,5,6,6,7,7-Dodecafluorooctanedioic acid
anthracen-2-ylamide hydroxyamide (21)
Using anthracen-2-ylamine (0.20 g; 1.03 mmol) as starting
material the product was obtained as a brownish solid according
to standard procedure 2.
Yield: 0.08 g (0.14 mmol); 14%; C22H12F12N2O3; M = 580.32
g/mol; 1H NMR (DMSO-d6): d (ppm) = 7.48–7.57 (m, 2H); 7.71–
7.77 (m, 1H); 8.05–8.28 (m, 3H); 8.48–8.62 (m, 3H); 9.94 (s, br,
1H); 11.51 (s, br, 1H); 12.38 (s, br, 1H); 19F NMR (DMSO-d6): d
(ppm) = À117.3–(À117.6) [m, 2F]; À118.1–(À118.4) [m, 2F];
À120.8–(À121.3) [m, 4F]; À121.5–(À121.7) [m, 2F]; À121.8–
(À122.1) [m, 2F]; MS (Àp ESI): m/z = 579 (100%) [MÀ1]À.
Yield: 1.2 g (14.75 mmol); 25%; C14H13N4; M = 237.28 g/mol; 1H
NMR (DMSO): d (ppm) = 5.39 (s, br, 2H); 6.56 (d, 2H); 6.85 (m, 2H);
7.41 (m, 2H); 7.56 (m, 2H); 7.95 (s, 1H); 13C NMR (DMSO): d
(ppm) = 112.44; 113.59; 127.17; 128.58; 128.75; 129.42; 130.52;
131.63; 138.42; 149.66; RF (diethylether): 0.58; MS (+p APCI):
m/z = 237 (100%) [MH+].
5.3.21. 2,2,3,3,4,4,5,5,6,6,7,7-Dodecafluorooctanedioic acid
hydroxyamide phenan-thren-9-ylamide (22)
Using phenanthren-9-ylamine (0.20 g; 1.03 mmol) as starting
material the product was obtained as a yellowish solid according
to standard procedure 2.
5.3.16. 2,2,3,3,4,4,5,5,6,6,7,7-Dodecafluorooctanedioic acid
hydroxyamide naph-thalen-1-ylamide (17)
Using naphthalen-1-ylamine (0.14 g; 0.98 mmol) as starting
material the product was obtained as a colorless solid according
to standard procedure 2.
Yield: 0.06 g (0.11 mmol); 11%; C18H10F12N2O3; M = 530.26
g/mol; 1H NMR (DMSO-d6): d (ppm) = 7.56–7.59 (m, 1H); 7.64–
7.73 (m, 3H); 7.83–7.91 (m, 1H); 8.04–8.14 (m, 2H); 10.02 (s, br,
1H); 10.62 (s, br, 1H); 12.44 (s, br, 1H); 19F NMR (DMSO-d6): d
(ppm) = À117.3–(À117.5) [m, 2F]; À118.1–(À118.4) [m, 2F];
À121.4–(À122.3) [m, 4F]; À122.6–(À123.1) [m, 2F]; À123.3–
(À123.8) [m, 2F]; MS (Àp APCI): m/z = 529 (100%) [MÀ1]À.
Yield: 0.03 g (0.05 mmol); 5%;
C22H12F12N2O3; M = 580.32
g/mol; 1H NMR (DMSO-d6): d (ppm) = 7.69–7.98 (m, 6H); 8.04–
8.11 (m, 1H); 8.89–8.99 (m, 2H); 9.97 (s, br, 1H); 11.63 (s, br,
1H); 12.46 (s, br, 1H); 19F NMR (DMSO-d6): d (ppm) = À117.3–
(À117.6) [m, 2F]; À118.3–(À118.6) [m, 2F]; À120.7–(À121.2) [m,
4F]; À121.3–(À121.6) [m, 2F]; À121.8–(À122.1) [m, 2F]; MS
(Àp ESI): m/z = 579 (100%) [MÀ1]À.
5.3.22. 2,2,3,3,4,4,5,5,6,6,7,7-Dodecafluorooctanedioic acid
anthracen-1-ylamide hydroxyamide (23)
Using anthracen-1-ylamine (0.20 g; 1.03 mmol) as starting
material the product was obtained as a brownish solid according
to standard procedure 2.
Yield: 0.10 g (0.17 mmol); 13%; C22H12F12N2O3; M = 580.32 g/
mol; 1H NMR (DMSO-d6): d (ppm) = 7.55–7.80 (m, 4H); 8.10–8.23
(m, 3H); 8.49 (s, 1H); 8.78 (s, 1H); 10.03 (s, br, 1H); 11.75 (s, br,
1H); 12.48 (s, br, 1H); 19F NMR (DMSO-d6): d (ppm) = À117.3–
(À117.6) [m, 2F]; À118.3–(À118.5) [m, 2F]; 120.7–(À121.1) [m,
4F]; À121.6–(À121.8) [m, 2F]; À121.9–(À122.2) [m, 2F]; MS (Àp
ESI): m/z = 579 (100%) [MÀ1]À.
5.3.17. 2,2,3,3,4,4,5,5,6,6,7,7-Dodecafluorooctanedioic acid
biphenyl-3-ylamide hydroxyamide (18)
Using biphenyl-3-ylamine (0.17 g; 1.00 mmol) as starting mate-
rial the product was obtained as a beige colored solid according to
standard procedure 2.
Yield: 0.08 g (0.14 mmol); 14%; C20H12F12N2O3; M = 556.30
g/mol; 1H NMR (DMSO-d6): d (ppm) = 7.39–7.43 (m, 1H); 7.45–
7.58 (m, 4H); 7.63–7.73 (m, 3H); 7.94–7.98 (m, 1H); 9.94 (s, br,
1H); 11.35 (s, br, 1H); 12.37 (s, br, 1H); 19F NMR (DMSO-d6): d
(ppm) = À116.9–(À117.2) [m, 2F]; À117.6–(À117.9) [m, 2F];
À120.0–(À120.4) [m, 4F]; À120.8–(À121.1) [m, 2F]; 121.2–
(À121.5) [m, 2F]; MS (Àp ESI): m/z = 555 (100%) [MÀ1]À.
5.4. Molecular simulation
5.3.18. 2,2,3,3,4,4,5,5,6,6,7,7-Dodecafluorooctanedioic acid
hydroxyamide naph-thalen-2-ylamide (19)
Using naphthalen-2-ylamine (0.14 g; 0.98 mmol) as starting
material the product was obtained as a beige colored solid accord-
ing to standard procedure 2.
Yield: 0.08 g (0.15 mmol); 15%; C18H10F12N2O3; M = 530.26 g/
mol; 1H NMR (DMSO-d6): d (ppm) = 7.55–7.64 (m, 2H); 7.87–8.04
(m, 1H); 7.91–8.05 (m, 3H); 8.37–8.40 (m, 1H); 10.02 (s, br, 1H);
11.53 (s, br, 1H); 12.42 (s, br, 1H); 19F NMR (DMSO-d6): d
(ppm) = À116.8–(À117.2) [m, 2F]; À117.6–(À118.0) [m, 2F];
A homology model of HDAC1 was generated on the basis of the
recently solved structure of HDAC218 using SWISS-MODEL.19
Homology modeling was rather straightforward, because both se-
quences share 93.5% identity in a 367 aminoacid overlap. Docking
studies of SAHA and the representative compound 4 to the homol-
ogy model of HDAC1 and the X-ray structures of HDAH (PDB 1ZZ1)
and HDAC8 (PDB 1T67) were performed using SwissDock20 in the
accurate mode and allowing the side chains to be flexible within
5 Å of any atom of the corresponding ligand in its reference binding
mode. For further refinement of the complex structures MD-simu-

L. M. Henkes et al. / Bioorg. Med. Chem. 20 (2012) 985–995
995
lations were performed using Gromacs V. 4.0.7.21 with an imple-
mented AMBER-99 forcefield.22 The Particle Mesh Ewald method
was used for the treatments of long-range electrostatic interac-
tions. Starting with the docking structures the enzyme–drug com-
plexes were subjected to two steps of energy minimization in
vacuo using steepest descent and conjugate gradient method to re-
move bad van der Waals contacts. Then the protein was solvated in
a box of 594 nm3 volume filled with Tip4 water molecules23 and
with PBS solution (pH 6.3). Finally, lysis buffer (50 mM Tris–HCl
pH 8.5; 1 M NaCl; 1 mM MgCl2; 1 w/v Triton-X-100) was added
up to 5 times of the corresponding pellet volume. The solution
was incubated at 37 °C for 2 h.
Acknowledgements
Na+ and ClÀ ions corresponding to
a final concentration of
This work was supported by a Grant of the Bundesministerium
für Wirtschaft und Technologie of the Federal Republic ofGermany.
The excellent encouragement of Dr. Ute Jochem is gratefully
acknowledged.
250 mM NaCl. After another energy minimization of the solvated
complex position-restrained dynamics simulations were run to
soak the water and the inhibitor into the inhibitor-enzyme com-
plex (200 ps at 100 K, 60 ps at 200 K and 200 ps at 300 K). Finally,
a productive MD-simulation was run over 2 ns at 300 K. The data
were analyzed using VMD 1.8.7.24
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmc.2011.11.041.
5.5. Fluorogenic two step enzyme activity assay
According to the two-step-assay established by Wegener et al.25
the protein assay was conducted in a total volume of 100
ll. In
References and notes
short, a fluorogenic substrate attached to a -amino-acetylated
e
lysine residue becomes deacetylated due to incubation with a cor-
responding HDAC or HDAH. Subsequently, trypsin digestion cleaves
the c-terminal ending of the lysine residue, releasing a fluorescent
7-amino-4-methylcoumarin (ex: 390 nm, em: 460 nm) group.
HDAH was prepared according to Hildmann et al.11, and human
recombinant HDACs were purchased from BPS BioSiences.
1. Rice, J. C.; Allis, C. D. Curr. Opin. Cell Biol. 2001, 13, 263.
2. de Ruijter, A. J.; van Gennip, A. H.; Caron, H. N.; Kemp, S.; van Kuilenburg, A. B.
Biochem. J. 2003, 370, 737.
3. Gregoretti, I. V.; Lee, Y. M.; Goodson, H. V. J. Mol. Biol. 2004, 338, 17.
4. Verdin, E.; Dequiedt, F.; Kasler, H. G. Trends Genet. 2003, 19, 286.
5. Aldana-Masangkay, G. I.; Sakamoto, K. M. J. Biomed. Biotechnol. 2011, 2011,
875824.
6. Witt, O.; Deubzer, H. E.; Milde, T.; Oehme, I. Cancer Lett. 2009, 277, 8.
7. Finnin, M. S.; Donigian, J. R.; Cohen, A.; Richon, V. M.; Rifkind, R. A.; Marks, P. A.;
Breslow, R.; Pavletich, N. P. Nature 1999, 401, 188.
8. Paris, M.; Porcelloni, M.; Binaschi, M.; Fattori, D. J. Med. Chem. 2008, 51, 1505.
9. Wu, R.; Lu, Z.; Cao, Z.; Zhang, Y. J. Am. Chem. Soc. 2011, 133, 6110.
10. Bradner, J. E.; West, N.; Grachan, M. L.; Greenberg, E. F.; Haggarty, S. J.;
Warnow, T.; Mazitschek, R. Nat. Chem. Biol. 2010, 6, 238.
11. Hildmann, C.; Ninkovic, M.; Dietrich, R.; Wegener, D.; Riester, D.;
Zimmermann, T.; Birch, O. M.; Bernegger, C.; Loidl, P.; Schwienhorst, A. J.
Bacteriol. 2004, 186, 2328.
12. Mai, A.; Esposito, M.; Bordella, G.; Massa, S. Org. Prep. Proced. Int. 2001, 33, 391.
13. Riester, D.; Wegener, D.; Hildmann, C.; Schwienhorst, A. Biochem. Biophys. Res.
Commun. 2004, 324, 1116.
14. Riester, D.; Hildmann, C.; Haus, P.; Galetovic, A.; Schober, A.; Schwienhorst, A.;
Meyer-Almes, F. J. Bioorg. Med. Chem. Lett. 2009, 19, 3651.
15. Hildmann, C.; Wegener, D.; Riester, D.; Hempel, R.; Schober, A.; Merana, J.;
Giurato, L.; Guccione, S.; Nielsen, T. K.; Ficner, R.; Schwienhorst, A. J. Biotechnol.
2006, 124, 258.
Boc-
concentration cf = 15
HDAC6 [cf = 0045 g/100
HDAC8 [cf = 0011 g/100
tute Boc- -Lys(trifluoroacetyl)-MCA. Substrates were applied at fi-
nal concentrations of 20 M in all enzyme assays. Ki-values were
calculated from IC50-values according to Cheng and Prusoff.26 The
Km-values of HDAH was 14
M15 and the Km-values of HDAC1, 6,
L
-Lys(acetyl)-MCA was a suitable substrate for HDAH [final
g/100 l], HDAC1 [cf = 0025 g/100 l] and
l]. HDAC7 [cf = 0002 g/100 l] and
l] favored the trifluoracetylated substi-
l
l
l
l
l
l
l
l
l
l
L
l
l
7, 8 were determined to be 78
lM, 99 lM, 61 lM and 30 lM,
respectively.
5.6. Dual parameter competition assay
The dual parameter competition assay introduced recently by
Riester et al.14 was used to measure binding of inhibitors to HDAH
at 21 °C. The assay principle exploits the concurrent change of fluo-
rescence anisotropy and FLT upon the reversible binding of the
fluorescent Atto700-hydroxamate ligand to the enzyme. The bind-
ing of an inhibitor is indicated by the reversal of the fluorescence
properties caused by an effective displacement of the ligand probe.
Final concentrations of 50 nM Atto-hydroxamate ligand and
16. Takai, N.; Narahara, H. J. Oncol. 2010, 2010, 458431.
17. Oger, F.; Lecorgne, A.; Sala, E.; Nardese, V.; Demay, F.; Chevance, S.; Desravines,
D. C.; Aleksandrova, N.; Le, G. R.; Lorenzi, S.; Beccari, A. R.; Barath, P.; Hart, D. J.;
Bondon, A.; Carettoni, D.; Simonneaux, G.; Salbert, G. J. Med. Chem. 2010, 53,
1937.
18. Bressi, J. C.; Jennings, A. J.; Skene, R.; Wu, Y.; Melkus, R.; De, J. R.; O’Connell, S.;
Grimshaw, C. E.; Navre, M.; Gangloff, A. R. Bioorg. Med. Chem. Lett. 2010, 20,
3142.
19. Kiefer, F.; Arnold, K.; Kunzli, M.; Bordoli, L.; Schwede, T. Nucleic Acids Res. 2009,
37, D387–D392.
2.5
lM HDAH were used.
20. Grosdidier, A.; Zoete, V.; Michielin, O. Nucleic Acids Res. 2011, 39, W270–
W277.
21. Hess, B.; Kutzner, C.; van der Spoel, D.; Lindahl, E. J. Chem. Theory Comput. 2008,
4, 435.
5.7. Cell culture
22. DePaul, A. J.; Thompson, E. J.; Patel, S. S.; Haldeman, K.; Sorin, E. J. Nucleic Acids
Res. 2010, 38, 4856.
23. Jorgensen, W. L.; Madura, J. D. Mol. Phys. 1985, 56, 1381.
24. Humphrey, W.; Dalke, A.; Schulten, K. J. Mol. Graphics Modell. 1996, 14,
33.
25. Wegener, D.; Hildmann, C.; Riester, D.; Schwienhorst, A. Anal. Biochem. 2003,
321, 202.
26. Cheng, Y.; Prusoff, W. H. Biochem. Pharmacol. 1973, 22, 3099.
K562 cancer cell line was established as a confluent cell culture
in RPMI 16040 medium supplemented with 2%
g/ml gentamycin and 10% FCS.
L-glutamine, 50
l
5.8. Cell lysis
Prior to lysis the K562 cells underwent centrifugation
(2000 rpm) for 5 min. Then the cells were collected and washed