Facile synthesis of colorimetric histone deacetylase substrates

Article abstract of DOI:10.1039/c2cc34422j

Here we report a simple procedure for generating colorimetric histone deacetylase (HDAC) substrates by solid-phase peptide synthesis based on racemization-free couplings of amino acid chlorides. We demonstrate the applicability of these substrates in HDAC assays.

Full text of DOI:10.1039/c2cc34422j

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
ChemComm
Cite this: Chem. Commun., 2012, 48, 9525–9527
www.rsc.org/chemcomm
COMMUNICATION
Facile synthesis of colorimetric histone deacetylase substratesw
a
a
Alexander Dose,z Jan Oliver Jost,z Antje C. Spieß,b Petra Henklein,c Michael Beyermannd
and Dirk Schwarzer*a
Received 20th June 2012, Accepted 3rd August 2012
DOI: 10.1039/c2cc34422j
Here we report a simple procedure for generating colorimetric
histone deacetylase (HDAC) substrates by solid-phase peptide
synthesis based on racemization-free couplings of amino acid
chlorides. We demonstrate the applicability of these substrates
in HDAC assays.
Lysine acetylation constitutes one of the most prominent
Scheme 1 Basic concept of commonly used HDAC assays.
protein modifications found in eukaryotic organisms. Initially
discovered in histone proteins, many more cellular targets of
this modification are known today.1 In the case of histones,
After the addition of trypsin, the fluorescence of released 3
directly reports the prior deacetylation reaction.11
which package DNA into chromatin, reversible lysine acetylation
regulates gene activity of adjacent DNA.2,3 Lysine acetylation is
This type of assay is widely used and has been commercialized,
but the synthesis of the required substrates is challenging.
catalyzed by histone acetyl-transferases (HATs) and erased by
two types of histone deacetylases (HDACs): Zn2+-dependent
Established protocols condense Lys(Ac)-fluorophore (or chromo-
HDACs and NAD+-dependent sirtuins.4–6 HDACs have
phore) conjugates as a single building block with the C-terminus
of protected peptides in solution.11,12 This procedure is both
gained considerable attention recently because they have been
inconvenient compared to stepwise solid-phase peptide synth-
identified as promising drug targets for a number of human
diseases including cancer.7 At present two HDAC inhibitors
esis (SPPS), and limiting with regard to the potential position
of the Lys(Ac)-chromophore unit, which can only be installed
(HDACis) have been approved for cancer therapy and it is
at the very C-terminus of the HDAC substrates. A reported
likely that more of such drugs will be added to the medicinal
alternative employs pairs of fluorophores and quenchers
repertoire in the near future. The development of HDACis
installed at the N- and C-termini of the HDAC substrate and
relies on sophisticated assays for monitoring HDAC activity in
a central Lys(Ac) residue.13 In this case alternative trypsin cleavage
sites at Lys and Arg residues, which make up 20–25% of all
histone amino acids, cause a fluorescence increase which is
not related to the deacetylation reaction. A simple strategy for
installing Lys(Ac) residues next to a reporting fluorophore or
chromophore at any given position within a peptide substrate by
means of SPPS would overcome the above discussed limitations.
Here we report a versatile procedure for the synthesis of
colorimetric HDAC substrates by SPPS using the Fmoc/tBu
strategy. We devised a colorimetric readout mechanism based
on the p-nitroaniline derivative 5-amino-2-nitro-benzoic acid
(5,2-ANB), which has been successfully applied to trypsin
assays. Consistent with other reports, we observed that this
moiety could be coupled to a resin-bound amino acid without
protection of the 5-amino group.14–16 This is due to the low
nucleophilicity of this amino group, which is also the major
obstacle to using 5,2-ANB in SPPS. Since acid halides are
known to form this amide bond, we turned to amino acid
chlorides, which are rarely used in SPPS due to their ‘over-
activated’ nature.15–17 Chlorides of 9-fluorenylmethyloxycarbonyl
(Fmoc) and other urethane-protected amino acids tend to
form oxazolones in the presence of a base.18 Oxazolones are
less reactive than their corresponding chlorides and their
a high-throughput format.8,9 Suitable HDAC substrates
should be easily accessible because recent reports indicate that
residues in the vicinity of the acetylation site can influence the
substrate recognition of HDACs.10 The most frequently used
type of assay couples the HDAC activity to that of the
protease trypsin (Scheme 1).11 The employed substrates contain
an acetylated Lys moiety (Lys(Ac)) bound to the aromatic
amine of a fluorophore (or chromophore) as in 1. In this
setting the fluorescence of the fluorophore is efficiently
quenched and lysine acetylation protects the adjacent amide
bond from proteolysis by trypsin. Adding HDACs converts 1 to
2 and facilitates the cleavage of the respective amide bond in 2.
a Interfaculty Institute of Biochemistry, University of Tuebingen,
Hoppe-Seyler-Str. 4, 72076 Tuebingen, Germany.
E-mail: dirk.schwarzer@uni-tuebingen.de; Fax: +49-7071-29-4518;
Tel: +49-7071-29-73344
b RWTH Aachen, Worringer Weg 1, 52056 Aachen, Germany
c Institut fur Biochemie, Charite´ - Universitatsmedizin Berlin,
Oudenarder Straße 16, 13347 Berlin, Germany
¨
¨
d Leibniz-Institut fur Molekulare Pharmakologie (FMP),
¨
Robert-Roessle-Str. 10, 13125 Berlin, Germany
w Electronic supplementary information (ESI) available: Experimental
procedures and further experimental data. See DOI: 10.1039/c2cc34422j
z These authors contributed equally to this work.
c
This journal is The Royal Society of Chemistry 2012
Chem. Commun., 2012, 48, 9525–9527 9525

View Article Online
Scheme 2 Synthesis of DNBS-Lys(Ac)-Cl.
formation might even be reversible, which raises concerns about
racemization of the Lys(Ac) moiety.19,20 In contrast to urethane-
based protection groups, N-arenesulfonyl protectants cannot
form oxazolones from amino acid halides.21 Furthermore,
the greater inductive effect of the sulfonyl moiety increases
the reactivity of the respective acid chlorides. Consequently,
we selected the 2,4-dinitrobenzenesulfonyl (DNBS) group as a
protectant for the Lys(Ac) building block.22 This protecting
group can be removed with nucleophilic thiols under mild
conditions, which are fully compatible with SPPS protocols.
The synthesis of DNBS-Lys(Ac)-OH began with commercially
available Lys(Ac)-OH 4, which was first subjected to transient
Na silylation by treatment with N,O-bis(trimethylsilyl)acetamide
(BSA) (Scheme 2).23 The advantages of this course of action were
two-fold: the silylation activated the Na amino group for the
subsequent installation of the DNBS group and also mitigated
the otherwise poor solubility of Lys(Ac)-OH 4 in organic solvents
(here acetonitrile). The addition of 2,4-dinitrobenzenesulfonyl
chloride (DNBS-Cl) converted silylated 4 into the desired
product DNBS-Lys(Ac)-OH 5, which was obtained in excellent
yield (ESIw). Subsequently, we established a SPPS protocol
based on in situ formed DNBS-Lys(Ac)-Cl 6 obtained by
treating 5 with thionyl chloride (Table S1, ESIw) and achieved
72% conversion when 6 was coupled to resin-bound 5,2-ANB
under optimized conditions (ESIw). Deprotection of the DNBS
group with thiophenolate completed the incorporation of
the Lys(Ac)-5,2-ANB unit and subsequent couplings were
performed under standard SPPS conditions.
Fig. 1 Deacetylation reactions monitored at 340 nm by HPLC.
(a) Products formed after Sir2.1-catalyzed deacetylation and sub-
sequent cleavage with trypsin. (b) Chromatogram of 7. (c) Chromato-
gram of the Sir2.1-catalyzed deacetylation reaction of 7 after 5 minutes
of reaction time. (d) Chromatogram of the deacetylation reaction after
tryptic digestion. Incubation time with trypsin was selected based on
quantitative conversion of 8 to 10.
Based on this scheme we synthesized the colorimetric
HDAC substrate p53-5,2-ANB (7) covering residues 378 to
385 of the tumor suppressor protein p53. Acetylated lysine was
introduced at the well-established modification site Lys382
directly upstream of 5,2-ANB (Fig. 1a).24 To ensure that the
synthesis procedure preserved the chiral integrity of the
coupled Lys(Ac) residue 7 was subjected to chiral amino acid
analysis, which confirmed an optical purity of 99.5% of the
L-enantiomer (identical with the starting material).
We continued to study the properties of 7 in deacetylation
assays and performed a quantitative comparison of 7 and
p53-AMC (1a), a commercially available p53-derived HDAC
substrate that uses 7-amino-4-methylcoumarin (AMC) for
fluorescence readouts of deacetylase activities (Fig. 1). In these
assays we used the human homolog of Sir2.1, SIRT1, which is
known to deacetylate 1a and analyzed the deacetylation
activity of 7 and 1a in parallel. Kinetic parameters were
extracted from the progress curves with the incremental
method taking product inhibition into account (ESIw).26
We observed similar Vmax/Km values of 0.0128 minꢀ1 and
0.0121 minꢀ1 for 7 and 1a, respectively, with comparable levels
of product inhibition (Fig. S3 and Table S2, ESIw). Based on
this, we concluded that 7 performs equally well in deacetyla-
tion assays than the commercially available 1a.
In the following, we tested 7 as an HDAC substrate (Fig. 1a
and Fig. S1, ESIw). Sir2.1 from C. elegans belongs to the
NAD+-dependent HDACs of class III, also referred to as
sirtuins, which are known to deacetylate modified Lys382 of
p53.25 We treated 100 mM of 7 with recombinant Sir2.1 (1 mM)
in the presence of NAD+. The reaction was stopped after
5 minutes by adding nicotinamide, a known inhibitor of
sirtuins. In a first instance we analyzed the deacetylation
reaction by HPLC. After 5 minutes, 44% of 7 was deacety-
lated to 8, and addition of trypsin converted 7 to 9 and
deacetylated 8 to the yellow product 10 (Fig. 1b–d). Next we
confirmed that the reaction could be monitored by optical
readouts at 405 nm, an optimal wavelength for 5,2-ANB in 10
(Fig. S2, ESIw).
SIRT1 and its homologs have gained considerable attention
because elevated cellular levels of these enzymes have
been found to induce longevity in a variety of organisms.27
Consequently, efforts have been put forth to screen for small
c
9526 Chem. Commun., 2012, 48, 9525–9527
This journal is The Royal Society of Chemistry 2012

View Article Online
Notes and references
1 C. Choudhary, C. Kumar, F. Gnad, M. L. Nielsen, M. Rehman,
T. C. Walther, J. V. Olsen and M. Mann, Science, 2009, 325,
834–840.
2 T. Jenuwein and C. D. Allis, Science, 2001, 293, 1074–1080.
3 T. Kouzarides, Cell (Cambridge, Mass.), 2007, 128, 693–705.
4 M. Biel, V. Wascholowski and A. Giannis, Angew. Chem., Int.
Ed.., 2005, 44, 3186–3216.
5 P. A. Cole, Nat. Chem. Biol., 2008, 4, 590–597.
6 D. Schwarzer, J. Pept. Sci., 2010, 16, 530–537.
7 P. Gallinari, S. Di Marco, P. Jones, M. Pallaoro and
C. Steinkuhler, Cell Res., 2007, 17, 195–211.
8 A. T. Hauser, M. Jung and M. Jung, Curr. Top. Med. Chem., 2009,
9, 227–234.
9 B. Heltweg, J. Trapp and M. Jung, Methods, 2005, 36, 332–337.
10 C. E. Berndsen and J. M. Denu, Curr. Opin. Struct. Biol., 2008, 18,
682–689.
11 D. Wegener, F. Wirsching, D. Riester and A. Schwienhorst, Chem.
Biol., 2003, 10, 61–68.
Fig. 2 Resveratrol activation of SIRT1 with 100 mM of p53-5,2-ANB (7),
p53-AMC (1a) and Boc-Lys(Ac)-AMC (1b).
12 M. T. Borra, B. C. Smith and J. M. Denu, J. Biol. Chem., 2005,
280, 17187–17195.
molecule activators of human SIRT1, which ultimately led to
the discovery of resveratrol.28 However, subsequent investigations
indicated that this effect depends on substrates containing
hydrophobic fluorophores like AMC, rendering the physiological
relevance of this effect questionable.12,29 We reasoned that 7 might
be less affected by its reporter group because 5,2-ANB is less
hydrophobic than AMC and fully imbedded into the peptide. We
analyzed the SIRT1-catalyzed deacetylation of 7 side by side with
1a and 1b, both of which are conjugated to AMC. As predicted
from the kinetic analysis the deacetylation rates of 7 and 1a were
similar, but the deacetylation of 1b, which does not contain any
residues flanking the acetylation site, was about 10-fold less effective
(Fig. 2). In the presence of 200 mM resveratrol we detected a
5–7-fold rate acceleration for the deacetylation of 1a and 1b. In
contrast the deacetylation of 7 showed no significant rate enhance-
ments upon addition of resveratrol. This observation supports the
notion that the direct activation of SIRT1 by resveratrol depends
on AMC and may not occur with physiological substrates.
13 B. E. Schultz, S. Misialek, J. S. Wu, J. Tang, M. T. Conn,
R. Tahilramani and L. Wong, Biochemistry, 2004, 43, 11083–11091.
14 K. Hojo, M. Maeda, S. Iguchi, T. Smith, H. Okamoto and
K. Kawasaki, Chem. Pharm. Bull. (Tokyo), 2000, 48, 1740–1744.
15 B. Spichalska, A. Lesner, M. Wysocka, M. Sledz, A. Legowska,
A. Jaskiewicz, H. Miecznikowska and K. Rolka, J. Pept. Sci.,
2008, 14, 917–923.
16 M. Wysocka, A. Lesner, G. Majkowska, A. Legowska, K. Guzow,
K. Rolka and W. Wiczk, Anal. Biochem., 2010, 399, 196–201.
17 L. A. Carpino, M. Beyermann, H. Wenschuh and M. Bienert,
Acc. Chem. Res., 1996, 29, 268–274.
18 L. A. Carpino, H. G. Chao, M. Beyermann and M. Bienert, J. Org.
Chem., 1991, 56, 2635–2642.
19 D. T. S. Rijkers, H. P. H. M. Adams, H. C. Hemker and
G. I. Tesser, Tetrahedron, 1995, 51, 11235–11250.
20 C. W. Smith and R. S. Rasmussen, J. Am. Chem. Soc., 1949, 71,
1080–1082.
21 L. A. Carpino, D. Ionescu, A. El-Faham, P. Henklein,
H. Wenschuh, M. Bienert and M. Beyermann, Tetrahedron Lett.,
1998, 39, 241–244.
22 T. Fukuyama, M. Cheung, C. K. Jow, Y. Hidai and T. Kan,
Tetrahedron Lett., 1997, 38, 5831–5834.
In conclusion we have established a simple and general synthesis
procedure for peptide based HDAC substrates. We have demon-
strated the applicability of these substrates for colorimetric read-
outs of HDAC activities and could show that these assays perform
as well as commercially available substrates but appear to be less
prone to artefacts. The convenience of the synthesis procedure
allows for a greater flexibility in the designing process of HDAC
substrates with respect to the amino acids flanking the acetylation
site. This feature might be advantageous for future investigations
into enzymatic deacetylation properties of the 3600 recently
identified lysine acetylation sites of the human proteome.1
23 J. R. Bowser, P. J. Williams and K. Kurz, J. Org. Chem., 1983, 48,
4111–4113.
24 W. Gu and R. G. Roeder, Cell (Cambridge, Mass.), 1997, 90,
595–606.
25 H. Vaziri, S. K. Dessain, E. Ng Eaton, S. I. Imai, R. A. Frye,
T. K. Pandita, L. Guarente and R. A. Weinberg, Cell (Cambridge,
Mass.), 2001, 107, 149–159.
26 C. Michalik, T. Schmidt, M. Zavrel, M. Ansorge-Schumacher,
A. Spiess and W. Marquardt, Chem. Eng. Sci., 2007, 62,
5592–5597.
27 L. Guarente and C. Kenyon, Nature, 2000, 408, 255–262.
28 K. T. Howitz, K. J. Bitterman, H. Y. Cohen, D. W. Lamming,
S. Lavu, J. G. Wood, R. E. Zipkin, P. Chung, A. Kisielewski,
L. L. Zhang, B. Scherer and D. A. Sinclair, Nature, 2003, 425,
191–196.
29 M. Kaeberlein, T. McDonagh, B. Heltweg, J. Hixon, E. A. Westman,
S. D. Caldwell, A. Napper, R. Curtis, P. S. DiStefano, S. Fields,
A. Bedalov and B. K. Kennedy, J. Biol. Chem., 2005, 280,
17038–17045.
We would like to thank Andrew Plested and Jakob Suckale
for critical comments on the manuscript and Wolfgang Fischle
for the kind gift of the Sir2.1 expression plasmid. This work
was supported by the Emmy-Noether programme of the
Deutsche Forschungsgemeinschaft (DFG) (SCHW 1163/3-1).
c
This journal is The Royal Society of Chemistry 2012
Chem. Commun., 2012, 48, 9525–9527 9527