Subtype selective substrates for histone deacetylases

Article abstract of DOI:10.1021/jm0497592

To probe the steric requirements for deacylation, we synthesized lysine-derived small molecule substrates and examined structure-reactivity relationships with various histone deacetylases. Rat liver, human HeLa, and human recombinant class I and II histon

Full text of DOI:10.1021/jm0497592

J . Med. Chem. 2004, 47, 5235-5243
5235
Su btyp e Selective Su bstr a tes for Histon e Dea cetyla ses
Birgit Heltweg,† Franck Dequiedt,‡ Brett L. Marshall,§ Carsten Brauch,† Minoru Yoshida,| Norikazu Nishino,
Eric Verdin,§ and Manfred J ung*,†
Department of Pharmaceutical and Medicinal Chemistry, Westfa¨lische Wilhelms-Universita¨t Mu¨nster, Hittorfstrasse 58-62,
48149 Mu¨nster, Germany, Molecular and Cellular Biology Unit, FUSAGx, Gembloux 5030, Belgium, Gladstone Institute of
Virology and Immunology, University of California, San Francisco, California 94141-9100, Chemical Genetics Laboratory,
RIKEN, Hirosawa 2-1, Wako-shi, Saitama 351-0198, J apan, and Department of Applied Chemistry, Kyushu Institute of
Technology, Tabata-ku, Kitakyushu-shi 804-8550, J apan
Received March 29, 2004
To probe the steric requirements for deacylation, we synthesized lysine-derived small molecule
substrates and examined structure-reactivity relationships with various histone deacetylases.
Rat liver, human HeLa, and human recombinant class I and II histone deacetylases (HDACs)
as well as human recombinant NAD+-dependent SIRT1 (class III enzyme) were used in these
studies. A benzyloxycarbonyl substituent on the R-amino group yielded the highest conversion
rates. Replacing the ꢀ-acetyl group with larger lipophilic acyl substituents led to a pronounced
decrease in conversion by class I and II enzymes; the class III enzyme displayed a greater
tolerance. Incubations with recombinant FLAG-tagged human HDACs 1, 3, and 6 showed a
distinct subtype selectivity among small molecule substrates. The subtype selectivity of HDAC
inhibitors could be predicted with these substrates and an easily obtainable mixture of HDAC
subtypes.
In tr od u ction
tors have so far failed to address this question. Struc-
ture-activity studies on tripeptide substrates for HDACs
are available; however, these substrates all contain an
acetyllysine group.6,7 Thus, a probing of the region
around the zinc ion using larger acyl substituents on
artificial substrates seems warranted.
Histone deacetylases (HDACs) are enzymes that
deacetylate histones and certain nonhistone proteins,
thereby altering their conformational states or activi-
ties.1 HDACs mediate transcriptional repression, and
the aberrant recruitment of deacetylase activity has
been suggested as a potential molecular mechanism
underlying certain types of leukemia. HDAC inhibitors
usually relieve transcriptional repression and result in
apoptosis or differentiation of cancer cells. Clinical
studies on HDAC inhibitors as new anticancer agents
are under way.2 Two classes of HDACs with 11 subtypes
are zinc-dependent amidohydrolases. A third class of
HDACs requires NAD+ for catalysis and is not sensitive
to class I and II inhibitors.3
Information on the subtype selectivity of available
inhibitors is limited, and the consequences of such
selectivity are unclear. No structural information on
mammalian class I or II HDACs is available. However,
the X-ray structure of a bacterial homologue has been
resolved4 and used to construct a homology model of
human HDAC1. This model showed a potential link
between certain biological activities and the different
binding modes of structurally similar inhibitors, thus
implying subtype selectivity.5 Structure-activity studies
of HDAC inhibitors may help to refine such models for
other subtypes. One open question is whether the deep
internal cavity “beyond” the zinc ion at the active site
is also present in mammalian HDACs. Available inhibi-
Previously, we introduced a small molecule substrate
for HDACs, fluorescent tert-butyloxycarbonyl (Boc)-(Ac)-
Lys-7-amino-4-methylcoumarin (AMC) (3a ), also termed
MAL,8 which can be used in various assay formats, such
as extraction, high-performance liquid chromatography
(HPLC), or plate reader9 measurements, and in a
homogeneous assay.10 We also introduced its benzyloxy-
carbonyl analogue Z-MAL (5a ), a substrate for human
HDACs that is also a small molecule substrate for
NAD+-dependent deacetylases.11 In the same study, 3a
had a moderate HDAC subtype selectivity. Here, we
present new analogues of 3a and 5a as probes for HDAC
structure and reactivity.
Resu lts
Syn th esis. For new substrate analogues, we prima-
rily modified the protecting group on the R-nitrogen and
the acyl substituent on the ꢀ-amino group. Benzyloxy-
carbonyl (Z, Cbz) and fluorenylmethyloxycarbonyl (Fmoc),
as well as the original Boc, were investigated as protect-
ing groups. Because the R-Z-analogue 5a was converted
faster than 3a ,11 the ꢀ-acyl group was varied starting
from 5a rather than from 3a . The acyl groups that we
studied were propionyl, butyryl, phenylacetyl, trifluo-
roacetyl, and also Boc in the urethanes 12b and 14.
Compound 14 is a rigid lysine analogue synthesized
from the racemic acid 13, which is commercially avail-
able. Additionally, the enantiomer 7 of 5a , the â-Z-
analogue 10, and compound 3b that has a trifluoro-
methyl instead of a methyl group on the coumarin
* To whom corresponding should be addressed. Department of
Pharmaceutical Sciences, University of Freiburg, Albertstrasse 25,
79104 Freiburg, Germany. Tel: +49-761-203-4896. Fax: +49-251-203-
6351. E-mail: manfred.jung@pharmazie.uni-freiburg.de.
† Westfa¨lische Wilhelms-Universita¨t Mu¨nster.
‡ FUSAGx.
§ University of California.
| RIKEN.
Kyushu Institute of Technology.
10.1021/jm0497592 CCC: $27.50 © 2004 American Chemical Society
Published on Web 09/08/2004

5236 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 21
Heltweg et al.
F igu r e 1. Synthesis of potential HDAC substrates.
moiety were synthesized. In general, coumarin was
linked to lysine derivatives with the POCl3/pyridine
activation procedure,12 as described in the original
synthesis of 3a .8 Mixed anhydride coupling was used
to synthesize 9 and 14. To introduce the acetyl and the
trifluoroacetyl moieties, the ꢀ-amino group was reacted
with the corresponding anhydride in pyridine. Acyl
chlorides and NaOH were used for the other acyl
substituents.13 In the synthesis of the â-analogue 10,
the coumarin was coupled first to an ꢀ-Boc-protected
â-Z-lysine and was ꢀ-acetylated in the last step after
cleavage of the Boc group (Figure 1).
Su bstr a te Rea ctivity. At a concentration of 100 µM,
none of the substrate analogues (3b, 5a -e, 7, 10, 12a ,b,
or 14) inhibited HDAC-mediated deacetylation of 3a ,
as monitored by HPLC.14 All potential substrates were
incubated at the same concentrations with a rat liver
HDAC preparation containing class I and II subtypes,
and their conversion was quantified by HPLC. All (S)-
ꢀ-acetyl-lysine-AMC derivatives (3a ,b, 5a , and 12a )
were good substrates for rat liver HDAC; 5a was the
best substrate (Table 1). The bulky R-Fmoc derivative
12a was not converted as readily as 5a . Similar results
were obtained with 3b when a trifluoromethyl group

Subtype Selective Substrates for Histone Deacetylases
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 21 5237
Ta ble 1. Conversion of Potential Substrates by Rat Liver
HDACa
% conversion
% conversion
substrate 1 h 2 h 3 h 6 h 24 h
substrate 1 h 2 h 3 h 6 h 24 h
3a
3b
5a
12a
5b
38 60 70 91 100
30 43 47 68 96
67 87 93 96 100
47 74 78 86 90
20 33 47 74 87
5c
5d
5e
7
0
0
1
0
12 21 48
0
0
0
16 41 57 54 64
14
0
9
0
3
a
Data are from duplicate experiments.
Ta ble 2. IC50 Determinations for HDAC Inhibitors
F igu r e 2. Conversion of substrates by recombinant human
hSIRT1. Data are from duplicate experiments. Compound 3a ,
clear bar; 3b, transversally lined bar; 5a , filled gray bar; 5b,
filled black bar; 5c, horizontally lined bar; 5d , vertically lined
bar. The variation between 5a and 3a and between 5a and
5b is considered to be significant (P < 0.05).
inhibitor
substrate
HDAC source
IC50 (nM)a
15
15
15
15
15
15
15
16a
16a
16a
16a
3a
3b
5a
5a
5b
5e
12a
3a
3b
5a
5a
rat liver
rat liver
rat liver
HeLa (Geneka)
rat liver
rat liver
rat liver
rat liver
rat liver
12.0 ( 1.2b
13.3 ( 1.4
26.3 ( 2.9
54.5 ( 4.0
20.7 ( 2.6
1526 ( 916
20.4 ( 1.5
216 ( 16b
153 ( 40
rat liver
HeLa (Geneka)
640 ( 33
1866 ( 134
a
b
Values are means ( SEM of duplicate experiments. Data are
from ref 9.
was introduced into the coumarin ring. The â-lysine
analogue 10 was not deacetylated by rat liver HDAC.
Even after 1 day of incubation, only 14% of enantiomer
7 of 5a was converted. The original substrates 3a and
5a were also studied using two commercially available
HeLa nuclear extracts as sources of human HDAC
actvity. In both cases, 5a was deacetylated to a much
greater extent.11 Thus, the variations of the ꢀ-acyl group
were performed with the R-Z-derivative. Chain exten-
sion to propionyl in 5b and to butyryl in 5c decreased
deacylation; the smaller 5b was the better substrate.
Because only 12% of the butyramide 5c was converted
within 3 h, longer chains were not studied. The bulkier
phenacetylamide 5d was not a substrate for rat liver
HDAC. Surprisingly, the R-Z-ꢀ-propionylamide 5b was
a better substrate for the HeLa nuclear extract than
the R-Boc-ꢀ-acetamide 3a (data not shown). The tri-
fluoroacetyl derivative 5e was also a substrate for rat
liver HDAC but was not as good as its acetyl analogue
5a , indicating that the electron-withdrawing effect of
the fluorine atoms does not increase deacetylation. The
carbamates 12b and 14 were not deacetylated by rat
liver HDAC. IC50 determinations of the inhibitors tri-
chostatin A (TSA, 15) and MD85 (16a ) revealed that
3a ,b, 5a ,b, and 12a can all serve as small molecule
substrates in HDAC assays (Table 2). The trifluoroacetyl
derivative 5e behaved differently as very large concen-
trations of 15 were needed to suppress substrate cleav-
age. Even at 2.8 µM, only 66% inhibition was obtained.
Sir tu in Activity. The substrates 3a ,b and 5a -d
were incubated with 8.75 U of the human recombinant
NAD+-dependent deacetylase hSIRT1 for 16 h at 37 °C
(Figure 2). Several substrates were converted in ratios
of about 15-45%. The phenacetyl compound 5d was
converted to about 25% but could not be cleaved by class
I or II HDACs from rat liver (Table 1). The trifluoro-
methyl analogue 3b was also converted to a greater
extent than 3a , which was not the case with rat liver
HDACs. Overall, the R-Z-ꢀ-acetyl analogue 5a was the
best substrate for all types of HDACs (see also ref 11).
F igu r e 3. Conversion of substrates by immunoprecipitated
human FLAG-tagged HDAC subtypes. WT: background activ-
ity with immunoprecipitation of wild-type cells. Data are from
five to eleven experiments. Compound 3a , clear bars; 3b, lined
bars; 5a , gray bars; 5b, black bars. For HDAC1, the variation
among the means was significant for 3a vs 5a (P < 0.001), 3b
vs 5a (P < 0.001), 5a vs 5b (P < 0.01), and 3b vs 5b (P <
0.05). For HDAC3, the variation among the means was not
significant. For HDAC6, the variation among the means was
significant for 3a vs 5b (P < 0.001), 3b vs 5b (P < 0.001), and
5a vs 5b (P < 0.001). For WT, the variation among the means
was not significant.
To determine if deacetylation of our lysine derivatives
is mediated by hSIRT1 as observed with histones or
peptidic substrates,15 we tried to detect O-acetyl ADP-
ribose by mass spectroscopy, but the standard buffer
interfered with the analysis. By using a modified Tris-
free buffer, we detected O-acetyl ADP-ribose in the
incubation of 5a with hSIRT1 (see Supporting Informa-
tion). These conditions, however, did not lead to a
similar detection of O-phenacetyl ADP-ribose in the
incubation of 5d . HPLC analysis of the reaction mix-
tures showed that the modified buffer decreased the
conversion of 5a from 45% after 16 h to 10% after 19 h
(see Supporting Information). No conversion of the less
reactive 5d was observed under these conditions (data
not shown).
Su btyp e Selectivity. Because 3a had a moderate
selectivity for HDAC6 vs HDAC1,11 we incubated the
most reactive substrates with immunoprecipitated hu-
man FLAG-tagged HDACs.16 As shown in Figure 3,
substrates 3a ,b and 5a ,b were converted after 3 h (5 h
for HDAC3). Compound 3a was deacetylated mostly by
HDAC6 and to some extent by HDAC1, but little
conversion by HDAC3 was observed even after longer
incubations. When the methyl group on the coumarin
ring in 3b was replaced with a trifluoromethyl group,
reactivity was lost with HDAC1 but was retained with
HDAC6. Compound 5a was deacetylated to a consider-

5238 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 21
Heltweg et al.
F igu r e 4. HDAC inhibitors.
Ta ble 3. Inhibitor Selectivity Screening with Substrates 3a ,b
and 5b
able extent by all three subtypes, but its homologue 5b
was selective for HDAC1 with some conversion by
HDAC3. Generally, HDAC3 activity was very low, a
finding confirmed by using tritium-labeled H4 peptide
as a substrate (data not shown), and no significant
preference of HDAC3 for one of the substrates was
observed.
IC50 (nM)a
selectivity
inhibitor
3a
3b
5b
(HDAC6/HDAC1)b
16b
16c
17a
17b
17c
78.2 ( 8.3 148 ( 14
256 ( 23 488 ( 165
349 ( 79 411 ( 20
137 ( 13 226 ( 16
3512 ( 999
241 ( 34
65.6 ( 8.4
21.2 ( 3.4
0.35
1.6
10
86
34
55.7 ( 7.6 80.4 ( 11.9 6.29 ( 1.02
Selectivity Scr een in g. In addition to the structural
information that may be gained from the reactivities,
we wanted to explore whether inhibitor selectivity could
be assessed by incubating a mixture of HDACs with
selective substrates. This would be much easier than
using immunoprecipitated subtypes and a nonselective
substrate. We compared the subtype selectivities of two
TSA-like inhibitors17 M344 (16b) and M360 (16c) and
three structurally unrelated cyclotetrapeptide inhibitors
(CHAPs18,19) (Figure 4). M344 (16b) was selective for
HDAC6 vs HDAC1 (IC50, 88 vs 249 nM). M360 (106 nM
for HDAC1; 165 nM for HDAC6) (16c) and CHAPs
17a -c were all selective for HDAC1. Incubations with
the HDAC1/HDAC3 selective substrate 5b and the
HDAC6 selective substrate 3b were performed, and the
results were compared with those obtained with sub-
strate 3a , which is converted by HDAC1 and HDAC6.
As shown with the homogeneous assay format,10 the
HDAC6 selective inhibitor 16b had a lower IC50 with
the HDAC6 selective substrate 3b; the opposite was
seen with the HDAC1 selective inhibitors 16c and
17a -c (Table 3).
a
b
Values are means ( SEM of duplicate experiments. Calcu-
lated by dividing the IC50 values on the immunoprecipitated
human enzymes using radioactively labeled histone peptides as a
substrate.
Discu ssion
Structure-reactivity studies on a set of synthetic
nonpeptidic HDAC substrates showed that conversion
rates and subtype selectivity are dependent on the
structure of the substrate. The ꢀ-acetyl-(S)-lysines were
good substrates. â-Lysine derivative 10 and (R)-lysine
derivative 7 were poor substrates but did not inhibit
the enzyme. Thus, the region of the enzyme that takes
up the lysine side chain is quite restricted to potential
branched chain substrates or inhibitors. Variations of
the R-protecting groups were less critical for substrate
acceptance by rat liver HDAC. HDAC1 seems to contain
a structural element distant from the zinc ion that can
distinguish between a methyl (in 3a ) and a trifluoro-
methyl (in 3b) substituent on the coumarin ring. This
element is not present in HDAC6, which converts both

Subtype Selective Substrates for Histone Deacetylases
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 21 5239
substrates to the same extent. A chain extension of the
ꢀ-acyl substituent is possible, but the butyrate 5c is
converted to a significant extent only within 1 day.
Increased lipophilic bulk, as in a phenylacetyl derivative
5d or in ꢀ-tert-butylcarbamates, abolished the ability to
serve as an HDAC substrate. These compounds are not
inhibitors, so we assume that they cannot access the
active site. Of course, the carbamates also have a
chemical reactivity that is different from that of the
amides. An ꢀ-trifluoroacetyl group in 5e also leads to
an HDAC substrate, but its conversion cannot be
inhibited completely with TSA at standard concentra-
tions (15). The reason for this is unclear. The propionyl
lysine 5b was selective for HDAC1 and HDAC3, sug-
gesting that those enzymes have a slightly bigger cavity
around the zinc ion than HDAC6.
of the newly discovered NAD+-dependent family of
HDACs. The sirtuin hSIRT1, which uses NAD+ to take
up the acyl group, has a wider range of acceptable
substrates. The same should be true for potential
inhibitors. Using mass spectrometry, we could show for
the ꢀ-acetyl derivative 5a that O-acetyl ADP-ribose is
formed with these small molecule substrates as a
consequence of the acyl transfer.
Finally, the use of the new subtype selective sub-
strates makes it possible to predict subtype selectivity
of HDAC inhibitors, at least semiquantitatively. With
further refinement, this approach could eliminate or
reduce the need to use immunoprecipitated HDAC
subtypes, simplifying lead discovery in that promising
class of compounds. HDAC6 has two catalytic domains,
one for histone deacetylation and one for tubulin
deacetylation.28 We do not know whether both domains
contribute to the conversion of our lysine coumarin
derivatives, but we propose that this is the case.
Concerning our HDAC inhibitors, we assume that these
compounds inhibit both catalytic domains of HDAC6 in
the same way as structurally similar inhibitors.25 These
assumptions are due to the absence of structural ele-
ments that favor binding to the tubulin deacetylase
domain as well as to the absence of an ortho-aminoben-
zamide element (favoring binding to the HDAC domain).
The HDAC6 selective M344 (16b) is a good inducer of
terminal cell differentiation in Friend leukemia cells17
and of γ-globin expression in a promotor assay and in
human erythroid cell cultures.29 Its more HDAC1 selec-
tive homologue M360 (16c) mainly exerts an antipro-
liferative activity in these assays. These findings war-
rant further investigation of HDAC inhibitor subtype
selectivity with respect to biological activity. The predic-
tive power of our subtype selective substrates for such
activities will be part of the investigations in that
direction.
The information from these experiments may be used
to refine homology models for HDAC subtypes. Such
models are the sole source of structural information for
the rational design of new selective inhibitors of this
therapeutically interesting class of potential drugs.
Concerning the existence of a remote internal cavity in
mammalian HDACs, our systematic studies support the
hypothesis that a large cavity is not present, raising
doubt about designing inhibitors that target to protein
regions beyond the zinc ion. However, results from a
docking study performed on the bacterial HDLP led the
authors to the opposite conclusion.20
The X-ray structure of human HDAC8 (Somoza, J .
R.; Skene, R. J .; Katz, B. A.; Mol, C.; Ho, J . D. et al.
Structural snapshots of human HDAC8 provide insights
into the class I histone deacetylases. Structure (Camb)
2004, 12, 1325-1334) was published while our manu-
script was in press. According to this work a large
internal cavity is indeed not present in human class I
enzymes as suggested by our work. Our results shed
light on the ongoing discussion21 about the mode of
action of HDAC inhibitors that contain a phenylenedi-
amine amide group, such as MS-275 (18),22-24 histacin
(19),25 and others.26 Although in vitro enzyme inhibition
by MS275 in the low micromolar range has been
reported,24 we could not reproduce such an inhibition
using our small molecule substrate or the commercially
available Biomol assay (data not shown). Recently,
using labeled histone and FLAG-tagged HDAC sub-
types, another group reported in vitro inhibition around
100 µM;27 we and others found cellular activity at
submicromolar concentrations.22 While attack on the
zinc ion at the active site is still possible and might be
stabilized by hydrogen bonding, which is unfavorable
for our butyramide or phenacetyl amides, it is feasible
that MS-275 and similar compounds bind at a different
Con clu sion
We used substrate analogues to assess the structural
requirements for compounds that access the catalytic
site of HDAC. For the region surrounding the zinc ion,
little variation seems to be possible for class I and II
HDACs, at least for lipophilic substituents. Fewer
restrictions were found with a member of the class III
family. These results demonstrate the potential for
developing bisubstrate sirtuin inhibitors. New subtype
selective substrates should be useful for refining homol-
ogy models of human HDACs; in the absence of struc-
tural data, such models are important tools for rational
drug design. These substrates will also allow inhibitors
to be tested for subtype selectivity with easily obtainable
extracts containing mixtures of different HDACs.
site (e.g., with a consecutive allosteric modulation21
which depends on certain cofactors.
)
Exp er im en ta l Section
In the published X-ray structure of HDAC8 with the
inhibitor TSA a second molecule TSA is bound closely
to the active site and maybe binding of an inhibitor to
this second binding site alone might result in HDAC
inhibition. Nevertheless, our conclusions drawn from the
longer acyl chain substrates must be viewed with some
caution.
Melting points are uncorrected. Elemental analysis was
performed on a Perkin-Elmer CHN-Analyzer. IR spectra were
recorded on a Shimadzu 470 in KBr. Wavenumbers are listed
in cm-1. 1H nuclear magnetic resonance (NMR) was done on a
Varian Mercury-400BB (400.3 MHz), and 13C NMR was done
on the same instrument (100.7 MHz). MS spectrometry for
compound identification was done on a Finnigan MAT 312 (EI).
Detection of O-acetyl ADP-ribose by MS spectrometry was
Several of our substrate analogues, among them the
phenylacetyl derivative 5d , which was not deacylated
by class I or II enzymes, are substrates for a member
performed by ESI on
a Waters-Micromass Quattro LCZ
equipped with a nanospray inlet. Flash chromatography was
performed with silica gel 60, 230-400 mesh (Merck). Pyridine

5240 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 21
Heltweg et al.
and tetrahydrofuran were dried over molecular sieves (3 Å).
Fluorescence spectra were recorded on a Shimadzu RF-540.
A Shimadzu RF 535 was used as the fluorescence detector for
HPLC and a LiChrosorb RP 18 5 µm column (125 mm × 3
mm, Knauer) or a Luna 5 µm Phenylhexyl column (250 mm
× 4.6 mm, Phenomenex) was used in the assay. For purity
testing of compounds by HPLC, a Kontron 335 UV detector
(254 nm) was used with the same two columns. TSA (15) was
from Sigma. Lysine derivatives were from Bachem. Other
chemicals were from Fluka and Aldrich. Compounds 3a ,8,30
5a ,11 16a -c,17,31,32 and 17a -c19 were prepared as described.
HeLa nuclear extracts were from Geneka (200 µg/40 µL,
catalog no. 100200) and Upstate (50 µg/25 µL, catalog no. 12-
309). hSIRT1 was from Biomol (catalog no. SE-239).
Hz), 1.75 (s, 3H), 1.61-1.23 (m, 15H). 13C NMR (DMSO-d6):
δ 172.61, 168.90, 158.62, 155.60, 154.73, 143.26, 139.18 (q, 2J C,F
1
) 32.03 Hz), 125.45, 121.69 (q, J C,F ) 275.4 Hz), 116.07,
114.36, 108.25, 106.23, 78.19, 55.35, 40.36, 31.17, 28.87, 28.18,
23.14, 22.60. MS (EI): m/z 243 (35), 126 (100), 84 (97). Anal.
(C23F3H28N3O6) C, H, N.
(S)-2-Ben zyloxycar bon ylam in o-6-pr opion ylam in o-h ex-
a n oic Acid (4b). Compound 4b was synthesized by method
A from Z-Lys (2.8 g, 10 mmol) and propionyl chloride (0.87
mL, 10 mmol). Yield: 0.89 g (26%), viscous oil.
(S)-2-Ben zyloxyca r b on yla m in o-6-b u t yr yla m in o-h ex-
a n oic Acid (4c). Compound 4c was synthesized by method
A from Z-Lys (2.8 g, 10 mmol) and butyryl chloride (1.05 mL,
10 mmol). Yield: 0.34 g (10%), viscous oil.
Syn th esis. Meth od A. Am id e F or m a tion w ith Acyl
Ch lor id es. To an ice-cold solution of N-R-benzyloxycarbonyl-
L-lysine (Z-Lys) in 1 M NaOH (1 mmol/mmol lysine) and H2O
(8 mL/mmol lysine), the acyl chloride (1 equiv) in anhydrous
tetrahydrofuran (1 mL/mmol acid chloride) and 1 M NaOH (1
mmol/mmol lysine) were separately added dropwise with
stirring over 20 min. The reaction mixture was saturated with
NaCl, cooled below 0 °C, and acidified to pH 1 with 2 M HCl.
The product was extracted with ethyl acetate, and the com-
bined organic layers were extracted with 5% NaHCO3 solution.
The bicarbonate solution was acidified and extracted three
times with ethyl acetate. The combined organic extracts were
washed with saturated brine and dried over Na2SO4, and the
solvent was evaporated.
Meth od B. Am id e F or m a tion w ith An h yd r id es. The
amine and acetic anhydride were dissolved in equimolar
amounts in dry pyridine. After stirring at room temperature
for 1 day, the mixture was acidified to pH 1 with 2 M HCl.
The product was extracted three times with ethyl acetate, and
the combined organic layers were extracted with 5% NaHCO3.
The bicarbonate solution was acidified and extracted three
times with ethyl acetate. The combined organic extracts were
washed with saturated brine and dried over Na2SO4, and the
solvent was evaporated.
Meth od C. Am id e F or m a tion w ith P h osp h or yl Ch lo-
r id e. The acid was mixed with equivalent amounts of AMC
(2a ) in dry pyridine (6 mL/mmol) at -15 °C. Phosphoryl
chloride (0.25 mL/mmol) was then added by syringe, resulting
in a orange-red solution. The mixture was stirred for 1 h at
-15 °C, poured into a 10-fold volume of ice/H2O, and extracted
three times with 50 mL of ethyl acetate. The combined organic
phase was washed consecutively with H2O, 2 M HCl, H2O, 5%
NaHCO3, H2O, and saturated brine (50 mL each). The organic
layer was dried over Na2SO4, and the solvent was evaporated.
Meth od D. Am id e F or m a tion w ith Mixed An h yd r id e.
To a solution of the acid in tetrahydrofuran (10 mL/mmol) was
added N-methyl-morpholine (NMM, 1 equiv) under nitrogen,
and the solution was stirred for 5 min. The solution was cooled
to -15 °C and stirred for another 5 min. Isobutyl chloroformate
(1 equiv) was added dropwise, and the mixture was stirred
for 10 min. AMC (2a ) (1 equiv) and NMM (2 equiv) were added,
and the suspension was stirred for 15 min at -15 °C and for
2 h at room temperature. The mixture was then poured into
50 mL of 2 M HCl solution and extracted three times with 50
mL of ethyl acetate. The combined organic phases were washed
consecutively with H2O, 2 M HCl, H2O, 5% NaHCO3, H2O (50
mL), and saturated brine. The organic layer was dried over
Na2SO4, and the solvent was evaporated.
(S)-2-Ben zyloxyca r bon yla m in o-6-p h en yla cetyla m in o-
h exa n oic Acid (4d ). Compound 4d was synthesized by
method A from Z-Lys (2.2 g, 7.85 mmol) and phenylacetyl
chloride (1.04 mL, 7.85 mmol). The resulting product was
chromatographed with ethyl acetate/n-hexane/acetic acid (2:
1:0.03). Yield: 1.97 g (63%), viscous oil.
(S)-2-Ben zyloxycar bon ylam in o-6-(2,2,2-tr iflu or oacetyl-
a m in o)h exa n oic Acid (4e). Compound 4e was synthesized
by method B from Z-Lys (2.2 g, 7.85 mmol) and trifluoroacetic
anhydride (1.11 mL, 7.85 mmol). Yield: 0.78 g (26%), viscous
oil.
(S)-[1-(4-Meth yl-2-oxo-2H-ch r om en -7-ylca r ba m oyl)-5-
pr opion ylam in o-pen tyl]car bam ic Acid Ben zyl Ester (5b).
Compound 5b was synthesized by method C from 4b (0.67 g,
2 mmol), 2a (0.35 g, 2 mmol), and POCl3 (0.4 mL). The
resulting product was chromatographed with ethyl acetate.
Yield: 0.44 g (45%); mp 161 °C. IR: 3313, 2939, 1695, 1620,
1
1585, 1526. H NMR (DMSO-d6): δ 10.47 (s, 1H), 7.76-7.34
(m, 10H), 6.25 (s, 1H), 5.02 (s, 2H), 4.18-4.00 (m, 1H), 3.00
(m, 2H), 2.39 (s, 3H), 2.01 (q, 2H, 3J ) 7.62 Hz), 1.74-1.52
(m, 2H), 1.47-1.19- (m, 4H), 0.94 (t, 3H, 3J ) 7.62 Hz). 13C
NMR (CDCl3): δ 172.34, 171.71, 159.76, 155.89, 153.38,
152.86, 142.00, 136.71, 128.15 (2C), 127.64, 127.57 (2C),
125.77, 115.06, 114.89, 112.15, 105.51, 65.43, 55.55, 38.17,
31.33, 28.92, 28.55, 23.13, 18.05, 10.09. MS (EI): m/z 447 (10),
107 (21), 91 (60), 79 (100). Anal. (C27H31N3O6) C, H, N.
(S)-[5-Bu tyr yla m in o-1-(4-m eth yl-2-oxo-2H-ch r om en -7-
ylca r b a m oyl)p en t yl]ca r b a m ic Acid Ben zyl E st er (5c).
Compound 5c was synthesized by method C from 4c (0.27 g,
0.77 mmol), 2a (0.14 g, 0.77 mmol), and POCl3 (0.15 mL). The
resulting product was chromatographed with ethyl acetate/n-
hexane (2:1). Yield: 0.15 g (38%); mp 156 °C. IR: 3306, 3091,
2960, 1688, 1619, 1584, 1529. 1H NMR (DMSO-d6): δ 10.47
(s, 1H), 7.76-7.29 (m, 10H), 6.25 (s, 1H), 5.01 (s, 2H), 4.18-
3
4.00 (m, 1H), 3.00 (m, 2H), 2.39 (s, 3H), 1.96 (q, 2H, J ) 7.42
3
Hz), 1.65-1.62 (m, 2H), 1.47-1.37 (m, 6H), 0.78 (t, 3H, J )
7.42 Hz). 13C NMR (DMSO-d6): δ 171.79, 171.56, 159.88,
155.97, 153.43, 153.00, 142.05, 136.76, 128.24 (2C), 127.72,
127.62 (2C), 125.85, 115.16, 114.98, 112.20, 105.60, 65.53,
55.61, 38.18, 37.48, 31.67, 28.96, 23.11, 18.85, 18.13, 13.75.
MS (EI): m/z 329 (65), 91 (65), 79 (100), 77 (73). Anal.
(C28H33N3O6) C, H, N.
(S)-[1-(4-Meth yl-2-oxo-2H-ch r om en -7-ylca r ba m oyl)-5-
p h en yla cetyla m in o-p en tyl]ca r ba m ic Acid Ben zyl Ester
(5d ). Compound 5d was synthesized by method C from 4d
(1.41 g, 3.54 mmol), 2a (0.62 g, 3.54 mmol), and POCl3 (0.6
mL). The resulting product was chromatographed with ethyl
acetate/methanol (10:1) and with dichloromethane/methanol
(20:1). Yield: 0.37 g (19%); mp 156 °C. IR: 3303, 2939, 1692,
1620, 1584, 1524. 1H NMR (CDCl3): δ 9.42 (s, 1H), 7.64-7.19
(S )-[5-Ace t yla m in o-1-(2-oxo-4-t r iflu or om e t h yl-2H -
ch r om en -7-ylca r ba m oyl)p en tyl]ca r ba m ic Acid ter t-Bu -
tyl Ester (3b). Compound 3b was synthesized by method C
from N-R-(Boc)-N-ꢀ-acetyl-lysine (1) (0.30 g, 1.04 mmol),
7-amino-4-trifluoro-methylcoumarin (AFC, 2b) (0.24 g, 1.04
mmol), and POCl3 (0.25 mL). The resulting product was
chromatographed with ethyl acetate/methanol (20:1). Yield:
0.28 g (54%); mp 121 °C. IR: 3346, 1734, 1704, 1619, 1580,
3
(m, 13H), 6.16 (s, 1H), 5.88 (d, 1H, J ) 7.62 Hz), 5.78 (t, 1H,
3J ) 5.66 Hz), 5.12-5.11 (m, 2H), 4.36-4.35 (m, 1H), 3.54 (s,
1H), 3.24-3.16 (m, 2H), 2.39 (s, 3H), 1.94-1.38 (m, 6H). 13C
NMR (CDCl3): δ 171.63, 170.90, 161.16, 156.56, 153.88,
152.41, 141.45, 136.02, 134.62, 129.38 (2C), 128.98 (2C),
128.50, 128.20 (2C), 127.95 (2C), 127.36, 125.02, 115.94,
115.77, 113.22, 107.19, 67.32, 55.45, 43.87, 38.70, 31.64, 28.78,
22.44, 18.80. MS (EI): m/z 91 (60), 79 (100), 77 (85). Anal.
(C32H33N3O6) C, H, N.
1
4
1526. H NMR (DMSO-d6): δ 10.58 (s, 1H), 7.90 (d, 1H, J )
1.70 Hz), 7.80 (t, 1H, 3J ) 5.29 Hz), 7.68 (d, 1H, 3J ) 7.36),
7.54 (dd, 1H, 3J ) 8.87 Hz, 4J ) 1.70 Hz), 7.16 (d, 1H, 3J )
3
7.36), 6.90 (s, 1H), 4.05-4.02 (m, 1H), 3.00 (d, 2H, J ) 5.48

Subtype Selective Substrates for Histone Deacetylases
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 21 5241
(S)-[1-(4-Meth yl-2-oxo-2H-ch r om en -7-ylca r ba m oyl)-5-
(2,2,2-tr iflu or oa cetyla m in o)p en tyl]ca r ba m ic Acid Ben -
zyl Ester (5e). Compound 5e was synthesized by method C
from 4e (0.64 g, 1.7 mmol), 2a (0.30 g, 1.7 mmol), and POCl3
(0.25 mL). The resulting product was chromatographed with
ethyl acetate/n-hexane (2:3) and with tetrahydrofuran/n-
hexane (1:1). Yield: 0.16 g (18%); mp °C. IR: 3309, 3095, 2947,
1705, 1620, 1584, 1528, 1182. 1H NMR (DMSO-d6): δ 10.47
(s, 1H), 9.39 (s, 1H), 7.75-7.29 (m, 9H), 6.23 (s, 1H), 5.01 (s,
2H), 4.13-4.10 (m, 1H), 3.15-3.14 (m, 2H), 2.38 (s, 3H), 1.66-
1.31 (m, 6H). 13C NMR (DMSO-d6): δ 171.69, 159.86, 155.97,
168.59, 159.77, 155.34, 153.38, 152.87, 142.24, 137.01, 128.08
(2C), 127.46, 127.33 (2C), 125.67, 114.98, 114.70 111.99,
105.40, 64.94, 48.03, 42.44, 31.88, 25.99, 22.65, 18.05. MS
(EI): m/z 95 (24), 91 (17), 79 (100), 74 (58). The purity was
more than 99%, tested by HPLC (two methods, UV detection).
(S)-[5-Acet yla m in o-1-(4-m et h yl-2-oxo-2H -ch r om en -7-
ylcar bam oyl)pen tyl]car bam ic Acid 9H-Flu or en -9-ylm eth -
yl Ester (12a ). Compound 12a was synthesized by method C
from N-R-(9-fluorenylmethoxycarbonyl)-N-ꢀ-acetyl-lysine (11a )
(0.50 g, 1.22 mmol), 2a (0.21 g, 1.22 mmol), and POCl3 (0.25
mL). The resulting product was chromatographed with ethyl
acetate/methanol (10:1). Yield: 0.17 g (25%); mp 197 °C. IR:
3282, 1734, 1664, 1623, 1526. 1H NMR (DMSO-d6): δ 10.48
(s, 1H), 7.88-7.30 (m, 13H), 6.25 (s, 1H), 4.28-4.23 (m, 4H),
3.00-2.99 (m, 2H), 2.39 (s, 3H), 1.75 (s, 3H), 1.70-1.39 (m,
6H). 13C NMR (DMSO-d6): δ 171.73, 168.62, 159.75, 155.87,
153.38, 152.86, 143.53 (2C), 142.00, 140.47 (2C), 127.45 (2C),
126.86, 125.78, 125.12 (2C), 119.94 (2C), 115.05, 114.90, 112.13
(2C), 105.51, 65.62, 55.42, 46.62, 38.29, 31.30, 28.90, 23.18,
22.66, 18.05. MS (EI): m/z 178 (100), 176 (22), 152 (16). The
purity was more than 99%, tested by HPLC (two methods, UV
detection).
2
155.96 (q, J C,F ) 35.1 Hz), 153.43, 152.90, 142.02, 136.73,
128.20 (2C), 127.68, 127.59 (2C), 125.76, 115.15, 114.97, 113.59
(q, 1J C,F ) 278.5 Hz), 112.21, 105.62, 65.53, 55.51, 39.71, 31.28,
28.03, 22.99, 18.13. MS (EI): m/z 271 (59), 180 (59), 175 (58),
91 (100), 79 (78), 77 (64). Purity was only 90%, tested by HPLC
(two methods, UV detection).
(R)-6-Acetylam in o-2-ben zyloxycar bon ylam in o-h exan o-
ic Acid (6). Compound 6 was synthesized by method B from
N-R-benzyloxycarbonyl-D-lysine (Z-D-Lys) (1 g, 3.57 mmol) and
acetic anhydride (0.34 mL, 3.57 mmol). Yield: 1.00 g (87%),
viscous oil.
(S)-[5-Boca m in o-1-(4-m et h yl-2-oxo-2H -ch r om en -7-yl-
ca r ba m oyl)p en tyl]ca r ba m ic Acid 9H-F lu or en -9-ylm eth yl
Ester (12b). Compound 12b was synthesized by method C
from N-R-(9-fluorenylmethoxycarbonyl)-N-ꢀ-(tert-butyloxy-car-
bonyl)lysine (R-Fmoc-ꢀ-OtBu-Lys, 11b) (0.50 g, 1.07 mmol), 2a
(0.19 g, 1.07 mmol), and POCl3 (0.2 mL). The resulting product
was chromatographed with ethyl acetate/n-hexane (3:2).
Yield: 0.20 g (30%); mp 180 °C. The purity was more than
99%, tested by HPLC (two methods, UV detection).
(R,S)-4-[2-(9H-Flu or en -9-ylm eth oxycar bon yl)-2-(4-m eth -
yl-2-oxo-2H-ch r om en -7-ylca r ba m oyl)eth yl]p ip er id in e-1-
ca r boxylic Acid ter t-Bu tyl Ester (14). Compound 14 was
synthesized by method D from N-R-(9-fluorenylmethoxyxar-
bonyl)-â-(1-butyloxycarbonyl-piperine-4-yl)-D,L-alanine (13) (0.50
g, 1 mmol), 2a (0.18 g, 1 mmol), NMM (0.33 mL, 3 mmol), and
isobutyl chloroformate (0.13 mL, 1 mmol). The resulting
product was chromatographed with ethyl acetate/n-hexane (1:
1). Yield: 0.17 g (26%); mp 140 °C. The purity was more than
99%, tested by HPLC (two methods, UV detection).
(R)-[5-Acet yla m in o-1-(4-m et h yl-2-oxo-2H -ch r om en -7-
ylca r b a m oyl)p en t yl]ca r b a m ic Acid Ben zyl E st er (7).
Compound 7 was synthesized by method C from 6 (0.94 g, 2.9
mmol), 2a (0.51 g, 2.9 mmol), and POCl3 (0.5 mL). The
resulting product was chromatographed with ethyl acetate.
Yield: 0.63 g (45%); mp 147 °C. IR: 3309, 2934, 2863, 1692,
1
4
1526. H NMR (DMSO-d6): δ 10.19 (s, 1H), 7.70 (d, 1H, J )
1.76 Hz), 7.61 (m, 1H), 7.50 d, 1H, 3J ) 8.60 Hz), 7.42 (dd,
3
4
3
1H, J ) 8.79 Hz, J ) 1.76 Hz), 7.26 (d, 1H, J ) 7.82 Hz),
7.14 (m, 5H), 6.18 (s, 1H), 5.07-4.88 (m, 2H), 4.24-4.06 (m,
1H), 3.11-2.84 (m, 2H), 2.32 (s, 3H), 1.98 (s, 3H), 1.79-1.38
(m, 6H). 13C NMR (DMSO-d6): δ 171.38, 168.93, 159.74,
155.73, 153.31, 152.12, 141.86, 136.35, 127.84 (2C), 127.35
(2C), 127.31, 124.89, 115.94, 114.75, 112.07, 105.78, 65.50,
55.42, 38.21, 31.37, 28.61, 22.83, 22.52, 18.07. MS (EI): m/z
126 (77), 91 (71), 84 (60), 79 (100), 77 (74). Anal. (C26H29N3O6)
C, H, N.
(S)-{4-Boc-a m in o-1-[(4-m e t h yl-2-oxo-2H -ch r om e n -7-
ylca r ba m oyl)m eth yl]bu tyl}ca r ba m ic Acid Ben zyl Ester
(9). Compound 9 was synthesized by method D from N-R-
benzyloxycarbonyl-N-ꢀ-(Boc)-â-lysine (8) (1.0 g, 2.63 mmol), 2a
(0.42 g, 2.63 mmol), isobutyl chloroformate (0.31 mL, 2.4
mmol), and NMM (0.69 mL, 7.2 mmol). The product was
chromatographed with ethyl acetate/n-hexane (2:1). Yield:
0.40 g (31%); mp 129 °C. IR: 3330, 2977, 1734, 1620, 1582,
1514, 1393, 1368, 1169. 1H NMR (CDCl3): δ 9.13 (s, 1H), 7.79
HDAC Assa ys. Ra t Liver En zym e. Rat liver HDAC was
purified by ammonium sulfate precipitation and chromatog-
raphy on Q-sepharose with an increasing gradient of sodium
chloride, as described.9 This preparation is also commercially
available (EMD Biosciences, Alexis).
HP LC Assa y. For the inhibition assay, the standard
substrate was 3a . For conversion assays, the different com-
pounds were used as substrates, and an internal standard,
7-hydroxycoumarin, was included for quantitation purposes.
The HPLC assay was performed on the LiChrosorb column
with acetonitrile/water (40/60 v/v) as the mobile phase at a
flow rate of 0.6 mL/min. The excitation wavelength was 330
nm (340 nm for 3b), and the emission wavelength was 390
nm (430 nm for 3b). The retention times were 4.05 (3a ), 10.26
(3b), 5.43 (5a ), 6.87 (5b), 9.42 (5c), 18.93 (5d ), 12.68 (5e), 5.43
(7), and 20.10 (12a ) min. To determine 12b and 14, we used a
Luna 5 µm Phenylhexyl column with methanol/water (90/10
v/v) as the mobile phase at a flow rate of 0.7 mL/min. The
excitation wavelength was 330 nm, and the emission wave-
length was 390 nm. Retention times were 9.95 (12b) and 12.70
(14) min. Results are from at least duplicate determinations.
To determine the resulting metabolites, we used a Luna 5 µm
Phenylhexyl column with acetonitrile/water/trifluoroacetic acid
(55/45/0.01 v/v/v) as the mobile phase at a flow rate of 0.5 mL/
min. The retention times were 4.01 min for the metabolite of
3a , 5.05 min for the metabolite of 3b, 3.99 min for the
metabolites of 5a -c,e and 7, and 5.40 min for the metabolite
of 12a . Stock solutions of compounds were made at 12.6 mM
in DMSO (3b, 5a ,c-e, 7, 10, and 12a ), 12.6 mM in ethanol
(3a ), or 12 mM in DMSO (5b, 12b, and 14) and further diluted
with enzyme buffer [1.4 mM NaH2PO4, 18.6 mM Na2HPO4,
pH 7.9, 0.25 mM EDTA, 10 mM NaCl, 10% (v/v) glycerol, and
10 mM 2-mercaptoethanol]. A stock solution was prepared
3
3
(d, 1H, J ) 8.29 Hz), 7.59 (s, 1H), 7.49 (d, 1H, J ) 8.77 Hz),
7.26-7.23 (s, 5H), 6.18 (s, 1H), 5.78 (d, 1H, 3J ) 8.53 Hz), 5.05
(s, 2H), 4.77 (s, 1H), 4.09-4.08 (m, 1H), 3.12 (m, 1H), 2.68 (d,
1H, J ) 5.12 Hz), 2.40 (s, 3H), 1.95-1.26 (m, 13H). 13C NMR
3
(CDCl3): δ 170.15, 161.72, 156.50, 156.33, 153.99 (2C), 142.12,
136.50, 128.59 (2C), 128.18, 128.01 (2C), 125.46, 116.17.
116.06, 113.15, 107.04, 79.55, 67.01, 49.06, 42.74, 40.46, 32.39,
28.81, 27.21, 19.07. MS (EI): m/z 437 (30), 307 (43), 286 (46),
285 (45), 263 (49) 175 (90), 91 (100), 79 (78), 77 (47).
(S)-{4-Acetyla m in o-1-[(4-m eth yl-2-oxo-2H-ch r om en -7-
ylca r ba m oyl)m eth yl]bu tyl}ca r ba m ic Acid Ben zyl Ester
(10). Compound 9 (0.32 g, 0.60 mmol) was dissolved in a
mixture of trifluoroacetic acid in dichloromethane (1:1, 4 mL).
After the mixture was stirred for 10 min, methanol was added,
and the solvents were evaporated. This step was repeated until
the smell of trifluoroacetic acid had disappeared. The resulting
amine was treated without further workup with acetic anhy-
dride (1 equiv) in pyridine, according to method B. The product
was precipitated after evaporation and redissolution in tet-
rahydrofuran with n-hexane. Yield: 0.26 g (93%); mp 198 °C.
IR: 3302, 1731, 1690, 1622, 1524. 1H NMR (DMSO-d6): δ
10.33 (s, 1H), 7.79 (t, 1H, 3J ) 5.27 Hz), 7.72 (d, 1H, 4J ) 1.95
Hz), 7.67 (d, 1H, 3J ) 8.6 Hz), 7.43 (dd, 3J ) 8.79 Hz, 4J )
1.76 Hz), 7.32-7.24 (m, 5H), 6.23 (s, 1H), 5.00-4.96 (m, 2H),
3.37 (m, 1H), 2.97-2.95 (m, 2H), 2.52-2.47 (m, 5H), 1.75 (s,
3H), 1.68-1.14 (m, 4H). 13C NMR (DMSO-d6): δ 169.48,

5242 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 21
Heltweg et al.
using an aliquot of 10 µL of the substrate (3a or compound to
test), 15 µL of a solution of the standard 7-hydroxycoumarin
in DMSO (6.3 mg/mL), and enzyme buffer to a total volume of
1 mL. For inhibitor screening, the stock solution of the
compound [12.6 mM in DMSO for 3b, 5a ,c-e, 7, 10, and 12a ;
12 mM in DMSO for 5b, 12b, 14 and MD85 (16a ); 3.3 mM in
DMSO for TSA (15)] was diluted with enzyme buffer in such
a fashion that a concentration was obtained that was 12-fold
higher than the highest test concentration desired in the assay.
(16 mM in methanol) in
a
ratio of 5/180/5 (v/v/v). The
fluorescence was measured at 330/390 nm. The amount of
substrate remaining in the samples containing the enzyme was
calculated relative to the amount of substrate in the positive
control (TSA) sample.
In h ibitor Selectivity. The subtype selectivity of inhibitors
16b,c was determined with a radioactive peptidic substrate
as described.34 Selectivity screening was performed with the
inhibitors CHAP1 (17a ), CHAP15 (17b), and CHAP31 (17c)
at final concentrations of 830, 280, 83, 28, 8.3, and 2.8 nM
and with M344 (16b) and M360 (16c) at final concentrations
of 10, 3.3, 1, 0.33, 0.1, and 0.033 µM and 3a ,b and 5b as
substrates in the homogeneous assay. The rat liver enzyme
(50 µL, prepared as described10 but with phosphate buffer
instead of Tris) was mixed with 10 µL of the buffer as blank,
with 5 µL of the substrate solution and 5 µL of the buffer as
negative control and with 5 µL of the substrate solution and 5
µL of the positive control solution (3.3 µM TSA) as a positive
control. Samples were prepared by mixing 50 µL of the enzyme
solution with 5 µL of each inhibitor solution and 5 µL of the
substrate solution. These mixtures were incubated at 37 °C
for 90 (3a ) and 180 (3b and 5b) min. The reaction was stopped
and analyzed as described above. The amount of remaining
substrate in the positive control with TSA was calculated
relative to the negative control without enzyme to give the
100% value. Partial conversion is related to that value.
The substrate/standard stock solution (10 µL) was added
to a mixture of 100 µL of rat enzyme preparation (at 4 °C)
and 10 µL of buffer or inhibitor solution. After 15 min at 4 °C,
the mixture was incubated at 37 °C for the desired time. The
reaction was stopped by adding 72 µL of 1 M HCl/0.4 M sodium
acetate and 800 µL of ethyl acetate, and the mixture was
centrifuged at 10000 rpm for 5 min. An aliquot (200 µL) of
the upper phase was taken, and the solvent was removed with
a stream of nitrogen. The residue was dissolved in 600 µL of
chromatography eluent, and 20 µL was injected by the au-
tosampler onto the HPLC system. The amount of remaining
substrate was calculated relative to the substrate control
without enzyme as the quotient of the peak area of the
substrate divided by the peak area of the internal standard.
For the determination of the metabolites, the incubation was
stopped by adding 1 mL of acetonitrile instead of 72 µL of 1 M
HCl/0.4 M sodium acetate, and the mixture was centrifuged
as above. The supernatant was removed, and 20 µL was
injected by autosampler onto the HPLC system.
Ack n ow led gm en t. Funding by Circagen Pharma-
ceuticals is gratefully acknowledged by B.H. and M.J .
Very helpful suggestions for the modified sirtuin buffer
were made by Drs. M. D. J ackson and J . M. Denu,
Oregon Science and Health University. We thank Dr.
H. Luftmann, University of Mu¨nster, for measuring the
ESI mass spectra for the detection of O-acetyl ADP-
ribose.
h SIRT1 In cu ba tion a n d O-Acetyl ADP Ribose Detec-
tion . To assess reactivity with hSIRT1, we used a different
enzyme buffer (25 mM Tris-HCl, pH 8.0, 137 mM NaCl, 2.7
mM KCl, and 1 mM MgCl2) for all dilutions and incubations.
The substrate solutions were prepared as described with stock
solutions of 3a ,b, 5a -d , and the internal standard 7-hydroxy-
coumarin. A NAD+ stock solution (6 mM) was prepared in
water. The substrate solution (5 µL) was mixed with 2.5 µL of
hSIRT1, 5 µL of the NAD+ solution, and enzyme incubation
buffer to a total volume of 60 µL. The incubation at 37 °C was
stopped after 16 h by adding 36 µL of 1 M HCl/0.4 M sodium
acetate, extracted with 400 µL of ethyl acetate, and consecu-
tively treated as described above. For mass spectroscopy
analysis, the reaction mixtures (60 µL total volume) contained
500 µM of the substrate, 300 µM NAD+, 1 mM dithiothreitol,
20 mM pyridine, and 1 mM MgCl2. In the positive ion mode
in the ESI spectrum, the peak for O-acetyl ADP-ribose can be
found at 602. MS/MS analysis of this peak led to fragments
(M + 1), such as adenosine diphosphate (428), adenosine
monophosphate (349), O-acetyl ribose monophosphate (274),
ribose monophosphate (232), and adenine (132) (see Support-
ing Information; peaks were assigned as described33).
Hom ogen eou s Assa y, Recom bin a n t HDACs. The homo-
geneous assay was performed as reported.10 Aliquots (10 µL)
of stock solutions of 3a ,b and 5b (12.6 mM in DMSO) were
diluted in enzyme buffer [1.4 mM NaH2PO4, 18.6 mM Na2-
HPO4, pH 7.9, 0.25 mM EDTA, 10 mM NaCl, 10% (v/v)
glycerol, and 10 mM 2-mercaptoethanol] to a total volume of
1 mL and used as substrate solutions. Immunoprecipitated
human FLAG-tagged HDAC1, HDAC3, and HDAC634,35 were
prepared as described. M2 agarose (Sigma) was used at 25 µL/
mL lysate. Lysates of wild-type 293T cells served as a control.
The immunoprecipitated enzymes were washed in enzyme
buffer and used as the enzyme source. M2 agarose beads (2.5
µL) with precipitated enzymes were resuspended in 47.5 µL
of buffer and mixed with 10 µL of buffer as blank. Five
microliters of the substrate solution and 5 µL of the buffer were
added instead of 10 µL of buffer to measure a negative control
(no inhibition), respectively. Five microliters of the substrate
solution and 5 µL of TSA solution (3.3 µM) were added to
measure a positive control (100% inhibition). These mixtures
were incubated at 37 °C for 180 min (300 min for the
immunoprecipitated HDAC3). The reaction was stopped by
adding 190 µL of freshly prepared mixture of TSA (3.3 µM in
enzyme buffer), borate buffer (0.1 mM boric acid, adjusted with
1 M NaOH to pH 9.5), and naphthalene-2,3-dicarboxaldehyde
Su p p or tin g In for m a tion Ava ila ble: Additional spectral
data for compounds 4b-d , 6, 12b, and 14, mass spectra of
O-acetyl ADP ribose, and information on compound purity with
HPLC and elemental analysis data. This material is available
free of charge via the Internet at http://pubs.acs.org.
Refer en ces
(1) Grozinger, C. M.; Schreiber, S. L. Deacetylase enzymes: Biologi-
cal functions and the use of small-molecule inhibitors. Chem.
Biol. 2002, 9, 3-16.
(2) J ohnstone, R. W. Histone-deacetylase inhibitors: novel drugs
for the treatment of cancer. Nat. Rev. Drug Discovery 2002, 1,
287-299.
(3) Gray, S. G.; Ekstrom, T. J . The human histone deacetylase
family. Exp. Cell Res. 2001, 262, 75-83.
(4) Finnin, M. S.; Donigian, J . R.; Cohen, A.; Richon, V. M.; Rifkind,
R. A.; et al. Structures of a histone deacetylase homologue bound
to the TSA and SAHA inhibitors. Nature 1999, 401, 188-193.
(5) Remiszewski, S. W.; Sambuccetti, L. C.; Atadja, P.; Bair, K. W.;
Cornell, W. D.; et al. Inhibitors of human histone deacetylase:
synthesis, enzyme and cellular activity of straight chain hy-
droxamates. J . Med. Chem. 2002, 45, 753-757.
(6) Wegener, D.; Hildmann, C.; Riester, D.; Schwienhorst, A.
Improved fluorogenic histone deacetylase assay for high-
throughput screening applications. Anal. Biochem. 2003, 321,
202-208.
(7) Wegener, D.; Wirsching, F.; Riester, D.; Schwienhorst, A. A
fluorogenic histone deacetylase assay well suited for high-
throughput activity screening. Chem. Biol. 2003, 10, 61-68.
(8) Hoffmann, K.; Brosch, G.; Loidl, P.; J ung, M. A nonisotopic assay
for histone deacetylase activity. Nucleic Acids Res. 1999, 27,
2057-2058.
(9) Heltweg, B.; J ung, M. A micro plate reader based nonisotopic
histone deacetylase activity assay. Anal. Biochem. 2002, 302,
175-183.
(10) Heltweg, B.; J ung, M. A homogeneous nonisotopic assay for
histone deacetylase activity. J . Biomol. Screening 2003, 8, 89-
95.
(11) Heltweg, B.; Dequidt, F.; Verdin, E.; J ung, M. A nonisotopic
substrate for assaying both human zinc and NAD+-dependent
histone deacetylases. Anal. Biochem. 2003, 319, 42-48.

Subtype Selective Substrates for Histone Deacetylases
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 21 5243
(12) Rijkers, D. T. S.; Adams, H. P. H. M.; Hemker, H. C.; Tesser, G.
I. A convenient synthesis of amino acid p-nitroanilides- synthons
in the synthesis of protease substrates. Tetrahedron 1995, 51,
11235-11250.
(13) Bernardi, L.; De Castiglione, R.; Fregnan, G. B.; Gla¨sser, A. H.
An experimental approach to long-lasting hypotensive eledoisin-
like peptides. J . Pharm. Pharmacol. 1967, 19, 95-101.
(14) Hoffmann, K.; Heltweg, B.; J ung, M. Improvement and valida-
tion of the fluorescence-based histone deacetylase assay using
an internal standard. Arch. Pharm. Pharm. Med. Chem. 2001,
334, 248-252.
(15) J ackson, M. D.; Schmidt, M. T.; Oppenheimer, N. J .; Denu, J .
M. Mechanism of nicotinamide inhibition and transglycosidation
by Sir2 histone/protein deacetylases. J . Biol. Chem. 2003, 278,
50985-50998.
(16) Fischle, W.; Dequiedt, F.; Fillion, M.; Hendzel, M. J .; Voelter,
W.; et al. Human HDAC7 histone deacetylase activity is associ-
ated with HDAC3 in vivo. J . Biol. Chem. 2001, 20, 35826-35835.
(17) J ung, M.; Brosch, G.; Ko¨lle, D.; Scherf, H.; Gerha¨user, C.; et al.
Amide analogues of trichostatin A as inhibitors of histone
deacetylase and inducers of terminal cell differentiation. J . Med.
Chem. 1999, 42, 4669-4679.
(18) Furumai, R.; Komatsu, Y.; Nishino, N.; Khochbin, S.; Yoshida,
M.; et al. Potent histone deacetylase inhibitors built from
trichostatin A and cyclic tetrapeptide antibiotics including
trapoxin. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 87-92.
(19) Yoshida, M.; Furumai, R.; Komatsu, Y.; Nishino, N.; Horinouchi,
S. Histone deacetylase as a new target for cancer chemotherapy.
Cancer Chemother. Pharmacol. 2001, 48 (Suppl. 1), S20-S26.
(20) Wang, D. F.; Wiest, O.; Helquist, P.; Lan-Hargest, H. Y.; Wiech,
N. L. On the function of the 14 A long internal cavity of histone
deacetylase-like protein: Implications for the design of histone
deacetylase inhibitors. J . Med. Chem. 2004, 47, 3409-3417.
(21) Bouchain, G.; Delorme, D. Novel hydroxamate and anilide
derivatives as potent histone deacetylase inhibitors: Synthesis
and antiproliferative evaluation. Curr. Med. Chem. 2003, 10,
2359-2372.
(22) J aboin, J .; Wild, J .; Hamidi, H.; Khanna, C.; Kim, C. J .; et al.
MS-27-275, an inhibitor of histone deacetylase, has marked in
vitro and in vivo antitumor activity against pediatric solid
tumors. Cancer Res. 2002, 62, 6108-6115.
(23) Saito, A.; Yamashita, T.; Mariko, Y.; Nosaka, Y.; Tsuchiya, K.;
et al. A synthetic inhibitor of histone deacetylase, MS-27-275,
with marked in vivo antitumor activity against human tumors.
Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 4592-4597.
(24) Suzuki, T.; Ando, T.; Tsuchiya, K.; Fukazawa, N.; Saito, A.; et
al. Synthesis and histone deacetylase inhibitory activity of new
benzamide derivatives. J . Med. Chem. 1999, 42, 3001-3003.
(25) Wong, J . C.; Hong, R.; Schreiber, S. L. Structural biasing
elements for in-cell histone deacetylase paralog selectivity. J .
Am. Chem. Soc. 2003, 125, 5586-5587.
(26) Fournel, M.; Trachy-Bourget, M. C.; Yan, P. T.; Kalita, A.;
Bonfils, C.; et al. Sulfonamide anilides, a novel class of histone
deacetylase inhibitors, are antiproliferative against human
tumors. Cancer Res. 2002, 62, 4325-4330.
(27) Kraker, A. J .; Mizzen, C. A.; Hartl, B. G.; Miin, J .; Allis, C. D.;
et al. Modulation of histone acetylation by [4-(acetylamino)-N-
(2-amino-phenyl) benzamide] in HCT-8 colon carcinoma. Mol.
Cancer Ther. 2003, 2, 401-408.
(28) Haggarty, S. J .; Koeller, K. M.; Wong, J . C.; Grozinger, C. M.;
Schreiber, S. L. Domain-selective small-molecule inhibitor of
histone deacetylase 6 (HDAC6)-mediated tubulin deacetylation.
Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 4389-4394.
(29) Cao, H.; Stamatoyannopoulos, G.; J ung, M. Induction of human
gamma globin gene expression by histone deacetylase inhibitors.
Blood 2004, 103, 701-709.
(30) Hoffmann, K.; Brosch, G.; Loidl, P.; J ung, M. First nonradioac-
tive assay for in vitro screening of histone deacetylase inhibitors.
Pharmazie 2000, 55, 601-606.
(31) J ung, M.; Hoffmann, K.; Brosch, G.; Loidl, P. Analogues of
trichostatin A and trapoxin B as histone deacetylase inhibitors.
Bioorg. Med. Chem. Lett. 1997, 7, 1655-1658.
(32) Schmidt, K.; Gust, R.; J ung, M. Inhibitors of histone deacetylase
suppress the growth of MCF-7 breast cancer cells. Arch. Pharm.
Pharm. Med. Chem. 1999, 332, 353-357.
(33) J ackson, M. D.; Denu, J . M. Structural identification of 2′- and
3′-O-acetyl-ADP-ribose as novel metabolites derived from the
Sir2 family of beta-NAD+-dependent histone/protein deacetyl-
ases. J . Biol. Chem. 2002, 277, 18535-18544.
(34) Fischle, W.; Emiliani, S.; Hendzel, M. J .; Nagase, T.; Nomura,
N.; et al. A new family of human histone deacetylases related
to saccharomyces cerevisiae HDA1p. J . Biol. Chem. 1999, 274,
11713-11720.
(35) Fischle, W.; Dequiedt, F.; Hendzel, M. J .; Guenther, M. G.; Lazar,
M. A.; et al. Enzymatic activity associated with class II HDACs
is dependent on a multiprotein complex containing HDAC3 and
SMRT/N-CoR. Mol. Cell 2002, 9, 45-57.
J M0497592