Design, synthesis, and biological evaluation of a small-molecule inhibitor of the histone acetyltransferase Gcn5

Article abstract of DOI:10.1002/anie.200453879

HATs off! The development of the first cell-permeable small-molecule inhibitor of the human histone acetyltransferase (HAT) Gcn5 opens up new possibilities for understanding the histone code. Based on kinetic data and the proposed mechanism of the acetyla

Full text of DOI:10.1002/anie.200453879

Communications
Epigenetics
Design, Synthesis, and Biological Evaluation of a
Small-Molecule Inhibitor of the Histone
Acetyltransferase Gcn5**
Markus Biel, Androniki Kretsovali,*
Efthymia Karatzali, Joseph Papamatheakis, and
Athanassios Giannis*
Histone proteins are basic components of the eukaryotic
chromatin.[1] They contain
a DNA-interacting globular
domain and a more flexible N-terminal region,[2] which is a
target for several posttranslational modifications. These
include the acetylation of lysine residues and the methylation
of lysine and arginine residues as well as the phosphorylation
of serine hydroxyl functions and the attachment of an
ubiquitin group.[3] The “histone-code hypothesis”[4] is based
on the assumption that these modifications create a specific
substitution pattern on the histone tails. This pattern is
readable by regulatory proteins,which connect the histone
code with fundamental cellular processes like activation or
repression of transcription. To date,acetylation is the best-
studied histone modification. It has been shown that the
adjustment of a specific acetylation balance on the histone N-
termini within a particular gene region is the result of a highly
regulated interplay of selective histone acetyltransferases
(HATs) and histone deacetylases (HDACs).[5]
For decoding the histone code a carefully directed
influence on fine-tuning these processes through the develop-
ment of low-molecular-weight and cell-permeable inhibitors
is of extraordinary importance. Furthermore new inhibitors
may open up new possibilities for treatment of pathological
diseases like cancer.[5,6] In contrast to the HATs,several small-
molecule inhibitors of the HDACs are known,and some of
them are in clinical trials.[7] Recently,anacardic acid was
identified in a broad screening of plant extracts with
antitumor activity as the first small-molecule inhibitor of
the HAT p300.[8]
To date,many of the transcription factors described have
intrinsic HAT activity.[9] In agreement with their biological
function and based on sequence homologies,HATs can be
[*] Dr. A. Kretsovali, E. Karatzali, Prof. Dr. J. Papamatheakis
Institute of Molecular Biology and Biotechnology
FORTH, Vassilika Vouton
P.O. Box 1527, 71110 Heraklion, Crete (Greece)
Fax: (+30)2810-391-101
E-mail: kretsova@imbb.forth.gr
Dipl.-Chem. M. Biel, Prof. Dr. A. Giannis
University of Leipzig
Institute of Organic Chemistry
Johannisallee 29, 04103 Leipzig (Germany)
Fax: (+49)-341-973-6599
E-mail: giannis@chemie.uni-leipzig.de
[**] This work was supported by the Fonds der Chemischen Industrie.
A.K. and J.P. are grateful for funding GSRT/PENED 2002.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
3974
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/anie.200453879
Angew. Chem. Int. Ed. 2004, 43, 3974 –3976

Angewandte
Chemie
divided into three families: the GNAT,CBP/p300,and MYST
families.[9] In vivo,histone acetyltransferases are always
associated with large multiprotein complexes. In yeast,for
example,the histone acetyltransferase yGcn5 is part of the
multiprotein complexes Ada and SAGA,the substrate
selectivities of which differ from those of the free enzyme.[9]
Here,we describe the development and biological eval-
uation of the first small-molecule inhibitor of the human
histone acetyltransferase Gcn5,a prominent member of the
GNAT family with high preference for histone H3 as a
substrate. Moreover,we present our approach,from the first
working hypothesis,based on published kinetic studies and
mechanistic proposals,[9,10] to the establishment of the struc-
tural class of a-methylene-g-butyrolactones as inhibitors.
The g-butyrolactone scaffold is a recurrent structural
motif in many natural products.[11] Considering the wide
sphere of biological activity,the g-butyrolactone scaffold can
be designated as a privileged structure in nature. For this
reason we envisaged the development of completely new
HAT inhibitors using the class of g-butyrolactones (Fig-
ure 1b). The inspiration for choosing appropriate substituents
methoxybenzyl alcohol to give 7. Treatment with lithium
bis(trimethylsilyl)amide converted 7 into the corresponding
ester enolate,which reacted with an aliphatic aldehyde to give
the hydroxycarboxlic acid 8. Compound 8 was not isolated,
and the ring-closing reaction was performed in a mixture of
chloroform and ethanol. The cleavage of the 4-methoxyben-
zyl ester to give the final product was achieved under mild
conditions by heating with acetic acid in molten phenol
(Scheme 1).[17]
Scheme 1. Synthesis of the g-butyrolactones 1–5: a) 4-methoxybenzylic
alcohol, n-hexane, toluene, 608C, 36 h, 88%; b) LiHMDS, THF, À788C,
1 h; c) RCHO, THF, À788C, 12 h; d) CHCl3, EtOH, RT, 72 h; e) phenol,
HOAc, 608C, 3 h. LiHMDS=lithium hexamethyldisilazide.
Figure 1. a) Tetrahedral intermediate in the postulated catalytic mecha-
nism of Gcn5, and b) general structure of the inhibitors with the butyr-
olactone framework (for more details see text). R=alkyl chain.
Butyrolactones 1–5 were tested in an in vitro HAT assay,
using recombinant Gcn5[13] or CBP[14] and commercially
available core histone proteins. As shown in Figure 2,
compounds 3–5 lead to an inhibition of CBP,whereas only
at the positions 2 to 4 of the butyrolactone scaffold derived
from the requirements of the postulated catalytic mechanism
to the substrate.[10] According to the model of the induced fit,
it is assumed that only after the binding of acetyl-CoA to
Gcn5 is the binding pocket accessible for the histone protein,
and a ternary complex of enzyme,acetyl-CoA,and histone
can be formed. The highly conserved general base Glu173,
which is surrounded by several unpolar amino acids to
increase its pKa,deprotonates the e-amino function of the
histone lysine side chain and enables a nucleophilic attack on
the neighboring thioester function of acetyl-CoA (Figure 1a).
The resulting tetrahedral intermediate is stabilized by a
hydrogen bond to the backbone amide of Cys177 and
decomposes giving the general base Glu173,acetylated
histone,and CoASH.
We translated these facts into a simple working hypothesis
for the development of new HAT inhibitors: the desired
structure should present a possible hydrogen-bond acceptor
for the backbone amide of Cys177 and a polar group for
interaction with Glu173. Moreover,the molecule should
possess an aliphatic side chain analogous to that of lysine.
The synthesis of the g-butyrolactones 1–5[12] started with
the regioselective ring-opening of itaconic anhydride 6 with 4-
Figure 2. Inhibition of the HATs CBP and Gcn5 by compounds 3–5.
While butyrolactones 3–5 inhibit the HAT CBP, only compound 3 dis-
plays an inhibition against Gcn5. The control C contains 10 mg core
histone protein, 0.05 mCi [3H]-acetyl-CoA, and either 10 ng CBP or
100 ng GST-Gcn5. The inhibitor concentration for each experiment was
5 mm. After separation of the histone proteins by SDS gel electropho-
resis (12% SDS), the amount of acylated histones was quantified by
phosphor image analysis. H3: histone 3; H4: histone 4.
Angew. Chem. Int. Ed. 2004, 43, 3974 –3976
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3975

Communications
butyrolactone 3 inhibits the HAT Gcn5. For determination of
active side is unlikely. As a result,derivatization of g-
butyrolactone 3[17] at position 2 is a promising starting point
for future structure–activity relationship studies.
the IC50 values,the amount of enzymatically acetylated
histone protein was determined as a function of the inhibitor
concentration. As a result,compounds 3–5 show only a weak
inhibition of CBP (IC50 values: 3 = 0.5 mm, 4 and 5 = 1.7–
2 mm,data not shown),whereas butyrolactone 3 leads to an
inhibition of Gcn5 with an IC50 value of 100 mm (Figure 3). In
this context it is important to note that in the presence of
acetyl-CoA the Kd value for binding of histone H3 to Gcn5 or
PCAF is approximately 100 mm.[9a,15,16] The affinity of inhibit-
tor 3 to the Gcn5 enzyme is comparable to that of the natural
substrate H3 and provides an excellent starting point for the
study of structure–activity relationships.
Received: January 29,2004
Revised: April 18,2004 [Z53879]
Keywords: biological activity · epigenetics ·
.
histone acetyltransferase · histone code · inhibitors
[1] R. D. Kornberg,J. O. Thomas, Science 1974, 184,865 – 868.
[2] K. Luger,A. W. Mader,R. K. Richmond,D. F. Sargent,T. J.
Richmond, Nature 1997, 389,251 – 260.
[3] S. L. Berger, Curr. Opin. Genet. Dev. 2002, 12,142 – 148.
[4] a) B. D. Strahl,C. D. Allis, Nature 2000, 403,41 – 45; b) M. Izuka,
M. M. Smith, Curr. Opin. Genet. Dev. 2003, 13,154 – 160.
[5] D. Chung, Trends Mol. Med. 2002, 8,S10 – S11.
[6] a) S. L. Schreiber,B. E. Bernstein, Cell 2002, 111,771 – 778; b) P.
Workman, Curr. Opin. Pharmacol. 2001, 1,342 – 352.
[7] T. A. Miller,D. J. Witter,S. Belvedere, J. Med. Chem. 2003, 46,
5097 – 5116.
[8] K. Balasubramanyam,V. Swaminathan,A. Ranganathan,T. K.
Kundu, J. Biol. Chem. 2003, 278,19134 – 19140.
[9] a) S. Y. Roth,J. M. Denu,C. D. Allis, Annu. Rev. Biochem. 2001,
70,81 – 120; b) R. Marmorstein, J. Mol. Biol. 2001, 311,433 –
444; c) M. J. Carrozza,R. T. Utley,J. L. Workman,J. CôtØ,
Trends Genet. 2003, 19,321 – 329.
[10] R. C. Trievel,J. R. Rojas,D. E. Sterner,R. N. Venkataramani,L.
Wang,J. Zhou,C. D. Allis,S. L. Berger,R. Marmorstein,
Proc.
Natl. Acad. Sci. USA 1999, 96,8931 – 8936.
Figure 3. For determination of the IC50 value, butyrolactone 3 was incu-
bated in different concentrations (c) together with 100 ng Gcn5, 3 mg
core histone protein (Sigma), and 0.05 mCi [14C]-acetyl-CoA (Amer-
sham). After separation by SDS gel electrophoresis (12% SDS), the
amount of acetylated histone proteins was quantified. Each data point
represents the average of two independent measurements. EA=en-
zyme activity.
[11] a) H. M. R. Hoffmann,J. Rabe, Angew. Chem. 1985, 97,96 – 112;
Angew. Chem. Int. Ed. Engl. 1985, 24,94 – 110; b) A. G.
Gonzµlez,M. H. Silva,J. I. Padrón,F. LØon,E. Reyes,M.
lvarez-Mon,J. P. Pivel,J. Quintana,F. EstØvez,J. Bermejo,
Med. Chem. 2002, 45,2358 – 2361; c) F. P. Kuhajda,E. S. Pizer,
J. N. Li,N. S. Mani,G. L. Frehywot,C. A. Townsend, Proc. Natl.
Acad. Sci. USA 2000, 97,3450 – 3454; d) H. M. Ogmundsdottir,
J.
G. M. Zoega,S. R. Gissurarson,K. Ingolfsdottir,
J. Pharm.
Pharmacol. 1998, 50,107 – 115; e) A. Randazzo,C. Debitus,L.
Minale,P. P. Garcia,M. J. Alcaraz,M. Paya,L. Gomez-Paloma,
J. Nat. Prod. 1998, 61,571 – 575; f) A. E. Mitchell,J. Zheng,B. D.
Hammcock,M. Lo Bello,A. D. Jones, Biochemistry 1998, 37,
6752 – 6759; g) A. Huwe,R. Mazitschek,A. Giannis, Angew.
Chem. 2003, 115,2170 – 2187; Angew. Chem. Int. Ed. 2003, 42,
2122 – 2138.
The length of the aliphatic side chain at position 4 of the g-
butyrolactone seems to be decisive for the biological activity.
In following enzymatic experiments we could determine that
the inhibition shows no time dependence. This result speaks
for a nonirreversible inhibition of the enzyme,and variation
of the substitution pattern at position 2 may lead to further
optimization of compound 3.
The identification of small-molecule inhibitors of the
histone acetyltransferases is absolutely necessary for the
evaluation of biological processes connected with the model
of the histone code. On the basis of published kinetic studies
and postulated catalytic mechanisms we were able to develop
and identify the Gcn5 inhibitor 3. Furthermore,we could
determine a nonirreversible inhibition of Gcn5,and therefore
a Michael addition of nucleophilic groups of the enzymeꢀs
[12] R. M. Carlson,A. R. Oyler, J. Org. Chem. 1976, 41,4065 – 4069.
[13] E. Smith,J. Belote,R. Schlitz,X. Yang,P. Morre,S. Berger,Y.
Nakatani,C. Allis, Nucleic Acids Res. 1998, 26,2948 – 2954.
[14] J. Chrivia,R. Kwok,N. Lamb,M. Hagiwara,M. Montminty,R.
Goodman, Nature 1993, 365,855 – 859.
[15] K. G. Tanner,M. R. Langer,Y. Kim,J. M. Denu, Biochemistry
2000, 39,11961 – 11969.
[16] K. G. Tanner,M. R. Langer,Y. Kim,J. M. Denu, J. Biol. Chem.
2000, 275,22048 – 22055.
[17] The structure of compounds 1–5 was determined by NOE
experiments.
3976
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 3974 –3976