Exploration of the HDAC2 foot pocket: Synthesis and SAR of substituted N-(2-aminophenyl)benzamides

Article abstract of DOI:10.1016/j.bmcl.2010.03.091

A series of N-(2-amino-5-substituted phenyl)benzamides (3-21) were designed, synthesized and evaluated for their inhibition of HDAC2 and their cytotoxicity in HCT116 cancer cells. Multiple compounds from this series demonstrated time-dependent binding kin

Full text of DOI:10.1016/j.bmcl.2010.03.091

Bioorganic & Medicinal Chemistry Letters 20 (2010) 3142–3145
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Exploration of the HDAC2 foot pocket: Synthesis and SAR of substituted
N-(2-aminophenyl)benzamides
Jerome C. Bressi, Andy J. Jennings, Robert Skene, Yiqin Wu, Robert Melkus, Ron De Jong, Shawn O’Connell,
*
Charles E. Grimshaw, Marc Navre, Anthony R. Gangloff
Takeda San Diego, 10410 Science Center Drive, San Diego, CA 92121, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of N-(2-amino-5-substituted phenyl)benzamides (3–21) were designed, synthesized and evalu-
ated for their inhibition of HDAC2 and their cytotoxicity in HCT116 cancer cells. Multiple compounds
from this series demonstrated time-dependent binding kinetics that is rationalized using a co-complex
crystal structure of HDAC2 and N-(4-aminobiphenyl-3-yl)benzamide (6).
Received 29 January 2010
Revised 25 March 2010
Accepted 26 March 2010
Available online 30 March 2010
Ó 2010 Elsevier Ltd. All rights reserved.
Keywords:
Histone deacetylase inhibitors
HDAC
Oncology
The balance of reversible acetylation and deacetylation of his-
tones is critical to the regulation of chromatin structure, and it di-
rectly affects gene expression in eukaryotic organisms. Histone
deacetylases (HDACs), a class of Zn metalloenzymes, catalyze the
deacetylation of lysine residues in histones and facilitate the struc-
tural stabilization of nucleosomes necessary for the first level of
DNA organization.1,2 Histone acetyltransferase (HAT) maintains
the balance of deacetylation.3,4 Perturbations of this balance have
been linked to cancer and inhibition of HDACs results in the
expression of genes that produce terminal differentiation, growth
arrest and/or apoptosis in a variety of cancer cells.1,5 Inhibition of
HDACs has been identified as a mechanism for treating cancer,
and multiple small-molecule HDAC inhibitors are advancing in
clinical trials.6–9 In addition, HDAC inhibitors have proven useful
in the treatment or prevention of protozoal diseases10 and
psoriasis.11
Published small-molecule HDAC inhibitors typically contain a
bidentate chelator, or warhead, that interacts with the catalytic
Zn2+ that is nestled 8 Å deep in the protein active site. The two
most common chelators found in these molecules are hydroxa-
mates and anilidesas exemplified by SAHA12 and MS-275,9 respec-
tively (Fig. 1). Previously published crystal structures of HDLP13 (an
HDAC homolog) and HDAC814 demonstrate how hydroxamate
inhibitors are coordinated to the catalytic zinc as the tether sits
in a hydrophobic tunnel capped with a moiety that sits at the sol-
vent protein interface. However, a 14 Å long internal cavity or ‘foot
pocket’ adjacent to the catalytic zinc remains unoccupied. While
others have suggested inhibitors which could make use of the foot
pocket through modeling,15,16 no crystallographic evidence has
been presented in HDAC8 or HDLP which shows specific examples
accessing the foot pocket.
Analysis of HDAC2 crystal co-complexes17 obtained in our lab-
oratories with both hydroxamates and anilides revealed that the
anilides accessed the foot pocket adjacent to the zinc binding site
that, by its nature of atom connectivity, the hydroxamate could
not. To explore this binding region more thoroughly, we used
N-(2-aminophenyl)benzamide as the minimal scaffold required
to interact with the HDAC2 catalytic machinery. Inspection of the
HDAC2 crystal structure suggested that substitutions off of the
5-position of N-(2-aminophenyl)benzamides would provide the
appropriate vector for reaching towards interactions with the foot
pocket. We designed and synthesized a series of N-(2-amino-5-
substituted phenyl)benzamides (3–21) and evaluated them in an
HDAC2 enzyme assay and cellular assays.
The synthesis of N-(2-amino-5-substituted phenyl)benzamides
(Scheme 1) begins with Boc protection of the commercially avail-
able 4-bromo-2-nitroaniline (1). Reduction of the nitro group and
subsequent acylation with benzoylchloride provides tert-butyl 2-
benzamido-4-bromophenylcarbamate (2). Standard Suzuki cou-
pling conditions followed by removal of the Boc group provided
the desired compounds (4–21). Substituting with the appropriate
starting material, compound 3 was synthesized using standard
methodology.18
Benzamides were evaluated as inhibitors of recombinant hu-
man HDAC2 using a protease coupled end-point assay with Ac-Ly-
s(Ac)-AMC as a substrate. The reaction product was quantified by
* Corresponding author. Tel.: +1 858 731 3531; fax: +1 858 550 0526.
E-mail address: agangloff@takedasd.com (A.R. Gangloff).
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.03.091

J. C. Bressi et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3142–3145
3143
Figure 1. SAHA (hydroxamate chelator) and MS-275 (anilide chelator).
O
O
O2N
H2N
Br
a, b, c
d, e
HN
Br
HN
R
1
BocHN
H2N
2
5-21
Scheme 1. Reagents and conditions: (a) Boc2O, Et3N, DMF, rt, 16 h; (b) EtOH/AcOH, Zn (dust), reflux, 30 min; (c) PhC(O)Cl, Et3N, DMF, rt, 2 h; (d) R–B(OH)2, (Ph3P)4Pd, 2 M
Na2CO3, dioxane, reflux, 18 h; (e) DCM/TFA, rt, 1 h.
fluorescence at 480 nm with excitation wavelength at 370 nm. The
activity of the benzamides was evaluated separately after 1 h and
24 h pre-incubation of inhibitor with HDAC2 (Table 1). Most of
the benzamides inhibited HDAC2 in the pIC50 range of 4.6–6.1
without pre-incubation. There was, however, a notable time-
dependent inhibition upon pre-incubation of inhibitor with HDAC2
at 24 h resulting in increased inhibition for most compounds. In-
deed, compounds 5, 8, 12, 18 and 20 demonstrated a 21–42-fold
increase in enzymatic inhibition. Compounds with large R groups,
such as benzothiophenes 7 and 15, along with benzofuran 19 did
not display any difference in enzymatic potency between the two
assays. In contrast, no time-dependence was noted with hydroxa-
mate inhibitors such as SAHA.
Compounds were also evaluated for cytotoxicity in a human
HCT116 colon cancer cell line. EC50 determinations were calculated
from remaining cell viability after 72 h incubation using the MTS
calorimetric readout (Table 1). In general, compounds with enzy-
matic activity below 100 nM in the 24 h pre-incubation assay
exhibited some level of cellular activity. Compounds 5 and 8 dem-
onstrated EC50s of 1.1 and 2.1 lM, respectively.
Table 1
Enzymatic and cellular activity of substituted N-(2-aminophenyl)benzamides
b
c
a
Compd
R
HDAC2 inhibition 1 h-IC50
HDAC2 inhibition 24 h-IC50
D
Cytotoxicity HCT116 EC50
3
4
5
6
7
8
9
H
10
3.1
3Â
na
2-Thiazolyl
Phenyl
4-Vinylphenyl
2-Benzothiophenyl
3-Furanyl
5.8
0.90
5.0
10
0.81
3.9
0.86
5.0
1.8
7.2
200
10
27
3.4
24
10
1.4
40
0.39
0.027
0.80
10
0.039
0.27
0.14
3.8
0.043
3.1
21
10
13
1.3
0.60
10
0.032
3.8
0.072
14Â
33Â
6Â
—
31
1.1
na
na
21Â
14Â
6Â
1.3Â
42Â
2Â
10Â
—
2.1
16
>20
na
3.2
na
na
na
na
na
4-Pyridyl
10
11
12
13
14
15
16
17
18
19
20
21
SAHA
4-Methoxyphenyl
4-Acetylphenyl
2-Furanyl
3-Methylphenyl
3-Nitrophenyl
3-Benzothiophenyl
3-Acetylphenyl
3-Methoxyphenyl
4-Bromophenyl
2-Benzofuranyl
4-Fluorophenyl
4-Trifluoromethyl-phenyl
—
2
2Â
40Â
—
>20
na
3.1
na
42Â
10
0.076
—
2.0
a
b
c
Values in lM are means of three experiments; na = not active.
1 h-IC50 = 1 h reaction of benzamides with HDAC2 and substrate. Values in
24 h-IC50 = 24 h incubation of benzamides with HDAC2 followed by 1 h reaction with substrate. Values in lM are means of three experiments.
l
M are means of three experiments.

3144
J. C. Bressi et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3142–3145
why, in general, so few warheads with appreciable binding are
known (many weaker warheads rely upon additional interactions
around the surface of the protein and the top of the binding pocket).
To investigate the transition from the unbound solvated ligand
to the bound form, a simulation was carried out at the 6-31G* level
of theory using the package GAMESS-US.19 The unbound ligand
was also simulated with AM1/VAMP as for the full system and the
resulting conformation agreed very well with that found at the
6-31G* level. Figure 3 shows the comparison between the bound
(AM1) and unbound (6-31G*/AM1) conformations of the prototyp-
ical ligand used for this work. Of note are the positions of the
hydrogen atoms. In the bound form, the hydrogen atoms of the
ortho-NH2 group are involved in hydrogen bonds with the side-
chains of histidines 145 and 146, the nitrogen of the ortho-NH2
group chelates the zinc, and the carbonyl oxygen interacts with
both the tyrosine sidechain hydroxyl and the zinc. In the unbound
conformation, the ortho-NH2 group forms an internal hydrogen
bond with the carbonyl oxygen. It is clear that the number of com-
plimentary interactions formed upon binding to the HDAC active
site outweigh the dual penalties of ligand desolvation (approx
4.4 kcal/mol for the unbound conformation using the AM1/SCRF
method in VAMP) and the disruption of the internal hydrogen bond.
It is the latter which seems likely to be responsible for the time-
dependence observed. The gross conformation of the anilides does
not appear to change by a great deal upon binding to the HDAC ac-
tive site and this binding can occur with the internal hydrogen
bond intact (though the ligand will not be able to access the pocket
as deeply). However, a proportion of these transiently-bound mol-
ecules will experience the hydrogen translation described in Figure
3 and form a substantially more stable complex with the active
site. By Le Chatelier’s principle, over time these transiently-bound
forms will be converted to the tightly-bound, pseudo-irreversible
form until an equilibrium is reached. This equilibrium is responsi-
ble for the final IC50 observed for this class.
Figure 2. The X-ray crystal co-complex of compound 5 with HDAC2 showing the
catalytic residues, the zinc atom (orange) and the surface of the site in green.
Using compound 3 as a reference compound, one can go further to
explain the variable time-dependence with the choice of R group. As
the size of R increases, the more difficult it will be for compounds to ac-
cess the active site and undergo the described hydrogen bond transla-
tion. However, once bound, the reverse process will also be relatively
slow. Of course, a point is reached at which the R group size is simply
too large to allow acceptable access to the site and the affinities will
be reduced as a result, at least over the time-frames considered here
(e.g., compounds 7 and 15). Figure 2 illustrates how these larger mole-
cules will be at a disadvantage when accessing the active site deeply
due to the bend in the tube from the surface to the foot pocket.
In conclusion, a series of simple benzamides were designed and
synthesized to explore the ‘foot pocket’ area of the Zn2+ active site
of HDAC2. Some of these compounds inhibited the enzyme in the
low micromolar range and exhibited moderate micromolar cellular
potency in an HCT116 cancer cell line. Additionally, some of these
benzamides demonstrated a time-dependent binding event that
resulted in enzyme inhibition in the low nanomolar range after a
24-h incubation period with HDAC2. This time-dependent binging
phenomena may positively affect the in vivo potency of anilides
and merits further investigation.
Figure 3. (Left) The system investigated by semi-empirical methods. The truncated
anilide ligand is shown in light blue, the zinc as a silver sphere and the protein
carbon atoms in green. Polar interactions are indicated by the dashed, magenta,
lines. (Right) A comparison of the bound ligand conformation, with hydrogen atoms
optimized with AM1, and the unbound ligand, fully-optimized at the 6-31G* level.
The bound ligand conformation is shown in light blue. The carbon atoms of the
unbound ligand are yellow with the internal hydrogen bond indicated by the
dashed, magenta, line.
The HDAC2 active site consists of the catalytic machinery, a lipo-
philic ‘tube’ which leads from the surface to this machinery, and a
‘foot pocket’ immediately adjacent to the machinery. A zinc atom
is held by Asp181, His183, and Asp269. The lipophilic tube is formed
by Gly154, Phe155, His183, Phe210, and Leu276. Thefootpocket res-
idues are Tyr29, Met35, Phe114, and Leu144. This pocket contains
multiple water molecules in co-complexes which lack a sizeable R
group. Figure 2 shows the HDAC2 active site with 5 bound and illus-
trates the foot pocket occupied by the phenyl R group of 5.
In order to study the mechanism behind the time-dependence
observed, as well as suggest additional molecules for synthesis, a
1.46 Å co-complex structure was used as the basis for some semi-
empirical calculations. Figure 3 shows the system simulated at the
AM1 level with the program VAMP (version 8.0b; Clark, T. et al., Erlan-
gen). For this work, only the hydrogen atoms were optimized due to
the use of truncation and the availability of such high-resolution
structural data. The figure shows the final geometry achieved and
indicates the numerous polar interactions formed. It is interesting
tonotehowtheanilidenitrogenisa surrogatefor thewater molecule
required during the hydrolysis of acetylated lysine residues in the
natural processing by this enzyme. The interactions shown indicate
Acknowledgements
The authors would like to acknowledge Michael Organ and Deb-
asis Mallik from Total Synthesis for synthetic support and Lu Zheng
for analytical chemistry support.
References and notes
1. Kousarides, T.; Verdel, A.; Lemercier, C.; Seigneurin-Berny, D. Curr. Opin. Genet.
Dev. 1999, 9, 40.

J. C. Bressi et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3142–3145
3145
2. Strahl, B. D.; Allis, C. D. Nature 2000, 403, 41.
3. Eberharter, A.; Becker, P. B. EMBO Rep. 2002, 3, 224.
14. (a) 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. Structure 2004, 12, 1325; (b) Vannini, A.; Volpari, C.;
Filocamo, G.; Caroli Casavola, E.; Brunetti, M.; Renzoni, D.; Chakravarty, P.;
Paolini, C.; De Francesco, R.; Gallinari, P.; Steinckuhler, C.; Di Marco, S. Proc. Nat.
Acad. Sci. U.S.A. 2004, 101, 15064.
15. (a) Wang, D.-F.; Wiest, O.; Helquist, P.; Lan-Hargest, H.-Y.; Wiech, N. L. J. Med.
Chem. 2004, 47, 3409–3417; (b) Wang, D.-F.; Helquist, P.; Wiech, N. L.; Wiest,
O. J. Med. Chem. 2005, 48, 6936; (c) Bieliaskas, A. V.; Pflum, M. K. H. Chem. Soc.
Rev. 2008, 37, 1402; (d) Moradei, O. M.; Mallais, T. C.; Frechette, S.; Paquin, I.;
Tessier, P. E.; Leit, S. M.; Fournel, M.; Bonfils, C.; Trachy-Bourget, M.-C.; Liu, J.;
Yan, T. P.; Lu, A.-H.; Rahil, J.; Wang, J.; Lefebvre, S.; Li, Z.; Vaisburg, A. F.;
Besterman, J. M. J. Med. Chem. 2007, 50, 5543.
16. Methot, J. L.; Chakravarty, P. K.; Chenard, M.; Close, J.; Cruz, J. C.; Dahlberg, W.
K.; Fleming, J.; Hamblett, C. L.; Hamill, J. E.; Harrington, P.; Harsch, A.;
Heidebrecht, R.; Hughes, B.; Jung, J.; Kenific, C. M.; Kral, A. M.; Meinke, P. T.;
Middleton, R. E.; Ozerova, N.; Sloman, D. L.; Stanton, M. G.; Szewczak, A. A.;
Tyagarajan, S.; Witter, D. J.; Secrist, J. P.; Miller, T. A. Biorg. Med. Chem. Lett.
2008, 18, 973.
4. Grozinger, C. M.; Schreiber, S. L. Chem. Biol. 2002, 9, 3.
5. Smith, K. T.; Workman, J. L. Int. J. Biochem. Cell Biol. 2008, 41, 21.
6. (a) Marks, P. A.; Richon, V. M.; Breslow, R.; Rifkind, R. A. Curr. Opin. Oncol. 2001,
6, 477–483; (b) Marks, P.; Rifkind, R. A.; Richon, V. M.; Breslow, R.; Miller, T.;
Kelly, W. K. Nat. Rev. Cancer 2001, 3, 194; (c) Paris, M.; Porcelloni, M.; Binaschi,
M.; Fattori, D. J. Med. Chem. 2008, 51, 3330.
7. Meinke, P. T.; Liberator, P. Curr. Med. Chem. 2000, 8, 211.
8. Remiszeweski, S. W.; Sambucetti, L. C.; Atadja, P.; Baiir, K. W.; Cornell, W. D.;
Green, M. A.; Howell, K. L.; Jung, M.; Kown, P.; Trogani, N.; Walker, H. J. Med.
Chem. 2002, 45, 753.
9. (a) Saito, A.; Yamashita, T.; Mariko, Y.; Nosaka, Y.; Nakanishi, O. Proc. Natl. Acad.
Sci. U.S.A. 1999, 96, 4592; (b) Suzuki, T.; Ando, T.; Tsuchiya, K.; Fukazawa,
N.; Saito, A.; Mariko, Y.; Yamashita, T.; Nakanishi, O. J. Med. Chem. 1999,
42, 3001.
10. Mai, A.; Cerbara, I.; Valente, S.; Massa, S.; Walker, L. A.; Tekwani, B. L.
Antimicrob. Agents Chemother. 2004, 48, 1435.
11. McLaughlin, F.; La Thangue, N. B. Curr. Drug Targets Inflamm. Allergy 2004, 3,
213.
12. (a) Butler, L. M.; Agus, D. B.; Scher, H. I.; Higgins, B.; Rose, A.; Cordon-Cardo, C.;
Thaler, H. T.; Rifkind, R. A.; Marks, P. A.; Richon, V. M. Cancer Res. 2000, 60,
5165; (b) Madeleine, D.; Jenny, V. Expert Opin. Investig. Drugs 2007, 16, 1111–
1120; (c) Richon, V. M.; Garcia-Vargas, J.; Hardwick, J. S. Cancer Lett. 2009, 280,
201.
17. The crystal structure of compound 5 in HDAC2 has been deposited in the PDB
as 3MAX.
18. All compounds were characterized by 1H NMR and LC–MS, showing greater
>98% purity and were consistent with reported values.
19. Schmidt, M. W.; Baldridge, K. K.; Boatz, J. A.; Elbert, S. T.; Gordon, M. S.; Jensen,
J. H.; Koseki, S.; Matsunaga, N.; Nguyen, K. A.; Su, S. J.; Windus, T. L.; Dupuis,
M.; Montgomery, J. A. J. Comput. Chem. 1993, 14, 1347.
13. 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.