Complementary PCAF-Coenzyme a mutagenesis: Chemoenzymatic synthesis of a novel enlarged coenzyme a analogue and evaluation of its biological activity

Full text of DOI:10.1002/cbic.201000286

DOI: 10.1002/cbic.201000286
Complementary PCAF–Coenzyme A Mutagenesis: Chemoenzymatic
Synthesis of a Novel Enlarged Coenzyme A Analogue and Evaluation of Its
Biological Activity
Leang Khim,[a] Juan Han,[b] Lynsey Willetts,[a, b] Kevin Brady,[b] Pauline Gillece,[a] Ouannassa Rached,[a, b]
Neil R. Thomas,*[b] and Eleni Stylianou[a, c]
Discovered in 1996,[1] p300/CBP-associated factor (PCAF) is a
histone acetyltransferase, also known as lysine acetyltransfera-
se 2B (KAT2B; EC No. 2.3.1.48) with crucial roles in the regula-
tion of genes that underlie cancer, inflammation, learning and
memory.[2–9] PCAF catalyses the transfer of an acetyl group
from acetyl coenzyme A to the e-amino group of specific lysine
residues in both histone and non-histone proteins. The reverse
reaction is mediated by histone deacetylases (HDACs), which
remove acetyl group(s) from specific acetyl lysine residues
(Scheme 1). PCAF has been shown to be a transcriptional coac-
of these acetyltransferases has intensified.[12,13] First reported
by Kevan Shokat’s group at UCSF, a complementary enzyme–
coenzyme modification approach has proved successful in tar-
geting the direct phosphorylation substrates of protein kinases
including v-Src kinases and the extracellular signal-regulated
kinase 2 (ERK 2) by using enlarged ATP analogues.[14,15] Hither-
to, this approach has not been applied to histone-modifying
enzymes. We found after inspection of the high-resolution
crystal structure of PCAF with coenzyme A bound (Figure 1)[16]
that the binding site of PCAF contained key enzyme–acetyl
CoA contacts. The mutation of specific amino acids in this area
could potentially allow the protein to accommodate an en-
Scheme 1. Schematic representation of reactions catalysed by lysine histone
acetyltransferases (HAT) and lysine histone deacetylases (HDAC).
tivator protein that activates transcription through its intrinsic
histone acetyltransferase (HAT) activity and interaction with
many other transcription proteins.[2,10] It has recently also been
shown to be a substrate for the transcriptional methylase
SET7/9.[11]
Knowledge of the function of transcription acetylases is frag-
mentary due to the overlapping substrate specificities of these
enzymes. However, over the last decade, interest in identifying
synthetic molecules that can probe the structure and function
[a] Dr. L. Khim, Dr. L. Willetts, Dr. P. Gillece, O. Rached, Dr. E. Stylianou
University of Nottingham, School of Biomedical Sciences
Queen’s Medical Centre
Nottingham, NG7 2UH (UK)
[b] Dr. J. Han, Dr. L. Willetts, Dr. K. Brady, O. Rached, Prof. Dr. N. R. Thomas
University of Nottingham, School of Chemistry
Centre for Biomolecular Sciences
University Park, Nottingham, NG7 2RD (UK)
Fax: (+44)115-951-3564
E-mail: neil.thomas@nottingham.ac.uk
[c] Dr. E. Stylianou
Cleveland Clinic Foundation, Lerner Research Institute
Department of Pathobiology
Figure 1. A) Structure of the PCAF–CoA complex[34] with the Gln581 residue
highlighted. It is predicted that when Gln581 is mutated to Ala or Gly, it
should accommodate acetyl-3’’-O-benzyl-CoA (1). B) Schematic drawing of
interactions of PCAF amino acid residues with CoA in the binding pocket.
9500 Euclid Avenue, Cleveland, Ohio 44195 (USA)
Phone: (+1)216-445-7156
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cbic.201000286.
2100
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2010, 11, 2100 – 2103

larged coenzyme A structure. Hence, complementary, chemi-
cally synthesised, enlarged coenzymes and site-specifically mu-
tated enzymes that accept the enlarged cofactor could be
combined and used to determine the natural substrate portfo-
lio and hence function of PCAF.
to generate novel CoA analogues in high yields from pante-
theine derivatives by using PPAT and DPCK in conjunction with
pantothenate kinase.[37]
The majority of CoA analogues reported to date are modi-
fied at the thiol position, which is likely to decrease their cata-
lytic activity. We chose the 3’’-hydroxyl, as this position is suffi-
ciently far away from the reactive thiol to minimise any reduc-
tion in the ability of the enlarged CoA to act as an acetyl
donor and is close to several residues in the active site with
moderate to large side chains (Figure 1B.). In this paper, we
describe the chemoenzymatic preparation of 3’’-O-benzyl-
acetyl-coenzyme A, from readily available d-pantethine
through a seven-step reaction including a coupled enzymatic
reaction by PPAT and DPCK (Scheme 2). This is the first report
of the synthesis of this analogue and the evaluation of its
activity with the transcriptional acetylase KAT2B/PCAF.
Extra groups need to be appended to CoA in order to
render it specific to PCAF mutants and to prevent it from bind-
ing to the natural form of the enzyme or other acetyltransfer-
ases. We initially designed and evaluated an enlarged CoA
with a modified benzyl group: 3’’-O-benzyl-coenzyme A (com-
pound 1), which was identified on the basis of an examination
of the first sphere of residues that contact CoA in the KAT2B/
PCAF active site and taking into account the catalytic activity
of mutants that had been previously reported.[17,18] The PCAF
mutant proposed to be best suited to accommodating this an-
alogue was created by mutating Gln581 to alanine or glycine,
thus creating extra space close to the 3’’-hydroxyl that could
accommodate the additional benzyl group (Figure 1). The
benzyl group was chosen in this “proof-of-principle study” as it
has a rigid bulky structure that would not significantly change
the lipophilicity of coenzyme A.
2-O-Benzyl-4-O-phospho-d-pantethine (5) was prepared
chemically from the commercially available starting material d-
pantethine. Compound 2 was prepared by selectively protect-
ing the primary and secondary hydroxyl groups of d-pante-
thine with tert-butyldimethylsilyl (TBDMS) and benzyl groups,
respectively (90% overall yield). Removal of TBDMS to release
the primary hydroxyl group led to 2-O-benzyl-d-pantethine (3)
in a high yield (88%). Phosphorylation of 3 with di-tert-butyl
N,N’-diethyl phosphoramidite, activated by 1H-tetrazole,
Numerous coenzyme A analogues have been synthesised by
chemical or chemoenzymatic methods including seleno-CoA,[19]
oxy-CoA,[20] decysteamine-CoA,[21] desulfo-CoA,[22] succinyl-
CoA,[23] crotonyl-CoA,[24] bromoacetyl-CoA,[25] S-acetonyl-CoA,[26]
S-(3-oxobutyl)-CoA,[27] acetonyldethio-CoA,[28] methylmalonyl-
carba(dethia)-CoA,[29] keto-CoA[21] and guano-CoA.[30] The enzy-
matic synthesis of CoA analogues was first achieved in 1964
when Michelson utilised ribonuclease T2.[31] Drueckhammer and
Strauss have since, independently used the CoA biosynthetic
enzymes phosphopantetheine adenylyltransferase (PPAT) and
dephospho-coenzyme A kinase (DPCK) to synthesise a range of
analogues that are modified at the thiol position.[32–36] More re-
cently, a one-pot procedure has been developed by Nazi et al.
formed
2-O-benzyl-d-pantethine-1-O-di-tert-butylphosphite,
which was oxidised in situ with meta-chloroperbenzoic acid
(m-CPBA) to produce the di-tert-butyl phosphate 4 in a 50%
yield. The corresponding phosphate 5 was obtained quantita-
tively from 4 after the removal of the tert-butyl groups in acid
(>98%).
Due to a lack of success in synthesising 3’’-O-benzyl-coenzy-
me A from 5 by the procedure of Moffatt and Khorana[38]
through chemical coupling of the modified pantethine and ad-
Scheme 2. Synthesis of acetyl-3’’-O-benzyl-coenzyme A. Reagents and conditions: a) TBDMSCl, 4-dimethylaminopyridine (DMAP), dry DMF, RT, overnight, 95%;
b) i: potassium tert-butoxide (1m in THF), THF, À788C, 0.5 h, ii: benzyl bromide, RT, overnight, 90%; c) tetrabutyl ammonium fluoride, THF, RT, 4 h, 88%; d) i:
1H-tetrazole, di-tert-butyl N,N’-diethylphosphoramidite, THF, RT, 4 h; ii: m-CPBA (70%), À408C, 10 min, warm to RT, 2 h, 52%; e) HCl/dioxane (0.1m), 4 h,
100%; f) dithiothreitol (DTT), ATP, MgCl2·6H2O, PPAT (0.02 U), DPCK (0.40 U), Tris buffer (50 mm, pH 7.6), 378C, 2 d, 10%; g) i: R’=14CH3, 14CH3COONa, 2,6-di-
chlorobenzoic acid, THF, reflux, 30 min; R’=CH3, CH3COOH, THF, reflux, 30 min; ii: 1,1’-carbonyldiimidazole, THF, 658C, 1.5 h; iii: 6, buffer (1 mm DTT, 0.5m imi-
dazole, pH 7.0), RT, 30 min.
ChemBioChem 2010, 11, 2100 – 2103
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
2101

enosine-2’,3’-cyclic phosphate-5’-phosphomorpholidate, the
E. coli coenzyme A biosynthetic enzymes PPAT and DPCK were
employed in an alternative method to complete the synthesis
of 6. Analogue 6 was synthesised by coupling a modified pan-
tetheine fragment with ATP in a reaction catalysed by PPAT,
and then phosphorylating selectively the 3’ position of the re-
sulting dephosphocoenzyme A species with DPCK (step f in
Scheme 2). As anticipated, compound 5 was successfully con-
verted to 6. After purification by HPLC on a C-18 column ac-
cording to the method described elsewhere.[36,39] Isotopes 1a
and 1b were prepared according to the procedure reported
by Clough.[40] These enzyme catalysed reactions demonstrate
for the first time that DPCK and PPAT are able to tolerate large
hydrophobic groups on the hydroxyl of their normal substrate
in addition to modification at the sulfur position.[26–30]
Activity assays with wild-type and Q581A PCAF against
acetyl-3’’-O-benzyloxy-CoA indicated that 1b was not a selec-
tively preferred substrate of the mutant, but was also accepted
as a competent substrate by the wild-type enzyme. Therefore,
unfortunately 1b was not suitable for use as a probe for the
determination of authentic substrates of PCAF (Figure 2). The
Figure 3. Effect of acetyl-3’’-O-benzyl-CoA on PCAF activity. A) Analysis of
PCAF inhibition by coincubation with acetyl-3’’-O-benzyl-CoA. B) Relative ac-
tivity of PCAF inhibited by acetyl-3’’-O-benzyl-CoA given as a semi-log plot
to allow the IC50 to be calculated. The inhibition of PCAF activity by acetyl-
3’’-O-benzyl-CoA was investigated by incubation at 308C for 30 min of a
30 mL reaction mixture composed of 6 mL of assay buffer (50 mm Tris-HCl,
pH 8.0, 10% glycerol, 1 mm DTT, 10 mm sodium butyrate and 200 nm tri-
chostatin A), 100 ng wild-type PCAF, 10 mm [2-14C]acetyl-CoA, 0–2 mm acetyl-
3’’-O-benzyl-CoA and 10 mm histone H3 peptide (22 amino acids: ARTKQ-
TARKSTGGKAPRKQLAT).
Figure 2. Activity of wild-type and Q581A PCAF with acetyl 3’’-O-benzyl-CoA.
The catalytic activity of PCAF was determined in the presence or absence of
acetyl-CoA or acetyl-3’’-O-benzyl-CoA. The activity was measured at 308C for
60 min in a 30 mL reaction mixture containing 6 mL of assay buffer (50 mm
Tris-HCl, pH 8.0, 10% glycerol, 1 mm DTT, 10 mm sodium butyrate and
200 nm trichostatin A), 100 ng wild-type or Q581A PCAF, 10 mm radiolabelled
acetyl-CoA substrate and 10 mm histone H3 peptide (22 amino acids: ARTKQ-
TARKSTGGKAPRKQLAT). Lanes: 1: control (without enzyme), 2: wild-type
PCAF+acetyl-CoA, 3: wild-type PCAF+acetyl-3’’-O-benzyl-CoA, 4: Q581A
PCAF+acetyl-CoA, 5: Q581A PCAF+acetyl-3’’-O-benzyl-CoA.
likely explanation for this is that the extra benzyl group is not
sufficiently bulky to prevent binding to wild-type PCAF, where-
as the enlarged binding pocket created by the Q581A muta-
tion does not confer it with preferential binding to the en-
larged CoA. Hence wild-type PCAF interacts more efficiently
with the analogue. As a result of this, the enlarged analogue
of CoA was evaluated for its ability to inhibit the activity of
KAT2B/PCAF by competing with acetyl-CoA. More than 70% of
enzyme activity was inhibited when 0.25 mm of compound 1b
was incubated with KAT2B/PCAF and acetyl-CoA (10 mm). Com-
plete inhibition was observed with 2 mm 1b. The IC50 of 1b
was estimated to be 167 mm (Figure 3), and its Ki value was cal-
culated to be 169 mm from a Dixon plot (1/v vs [I]; Figure 4.
Note, additional data for the Dixon plot were not possible due
to the limited supply of [2-14C]acetyl 3’’-O-benzyl-coenzyme A).
For comparison, the Ki for CoA with PCAF has been reported
as 0.44 mm.[9]
Figure 4. Dixon plot for determining the inhibition of PCAF by [2-12C]acetyl-
3’’-O-benzyl-CoA. [2-14C]Acetyl-CoA was used as the substrate at concentra-
*
^
tions of 1 ( ) and 8 mm ( ). To determine the Ki value of acetyl-3’’-O-benzyl-
CoA, the reaction was performed at 308C for 30 min in a 30 mL reaction mix-
ture composed of 6 mL of assay buffer (50 mm Tris-HCl, pH 8.0, 10% glycerol,
1 mm DTT, 10 mm sodium butyrate and 200 nm trichostatin A), 100 ng wild-
type PCAF, [2-14C]acetyl-CoA (1 and 8 mm), acetyl-3’’-O-benzyl-CoA (0, 50 and
100 mm) and 10 mm histone H3 peptide (22 amino acids: ARTKQTARKSTGG-
KAPRKQLAT). Ki was determined by plotting the velocity against concentra-
tion of [2-12C]acetyl-3’’-O-benzyl-CoA at each concentration of [2-14C]acetyl-
CoA.
In summary, we have presented here an efficient synthesis
of a new CoA derivative, 3’’-O-benzyl-coenzyme A from a com-
mercially available source d-pantethine, by employing a che-
moenzymatic route involving PPAT and DPCK.
2102
www.chembiochem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2010, 11, 2100 – 2103

[3] F. Mateo, M. Vidal-Laliena, N. Canela, A. Zecchin, M. Martine-Balbas, N.
Agell, M. Giacca, M. J. Pujol, O. Bachs, Nucleic Acids Res. 2009, 37, 7072–
7084.
[4] G. Xenaki, T. Ontikatze, R. Rajendran, I. J. Stratford, C. Dive, M. Krstic-De-
monacos, C. Demonacos, Oncogene 2008, 27, 5785–5796.
[5] A. Imhof, X.-J. Yang, V. V. Ogryzko, Y. Nakatani, A. Wolffe, H. Ge, Curr.
Biol. 1999, 9, 689–692.
[6] V. Sartorelli, P. L. Puri, Y. Hamamori, V. Ogryzko, G. Chung, Y. Nakatani,
J. Y. Wang, L. Kedes, Mol. Cell 1999, 4, 725–734.
[7] T. Maurice, F. Duclot, J. Meunier, G. Naert, L. Givalois, J. Meffre, A. Cꢁlꢁri-
er, C. Jacquet, V. Copois, N. Mechti, K. Ozato, C. Gongora, Neuropsycho-
Pharmacol. 2008, 33, 1584–1602.
Unfortunately the benzylated CoA analogue proved to be a
substrate for both the wild-type PCAF and the mutant variant
that was specifically prepared to accommodate the CoA modi-
fication. In fact, our results show that the CoA analogue can
compete with the natural substrate for the native enzyme’s
active site, thus indicating that the substrate specificity of
PCAF enzymes is much lower than anticipated. The production
of further enlarged CoA analogues might allow the preparation
of inhibitors or substrates with specificity for a single CoA-de-
pendent enzyme that allows them to be used both as biologi-
cal tools and potentially as therapeutic agents.
[8] W. G. Deng, Y. Zhu, K. K. Wu, Blood 2004, 103, 2135–2142.
[9] K. G. Tanner, M. R. Langer, J. M. Denu, Biochemistry 2000, 39, 11961–
11969.
[10] A. Obrdlik, A. Kukalev, E. Louvet, A. K. Farrants, L. Caputo, P. Percipalle,
Mol. Cell. Biol. 2008, 28, 6342–6357.
[11] T. Masatsugu, K. Yamamoto, Biochem. Biophys. Res. Commun. 2009, 381,
22–26.
[12] P. A. Cole, Nat. Chem. Biol. 2008, 4, 590–597.
[13] Y.-Y. Yang, J. M. Ascano, H. C. Hang, J. Am. Chem. Soc. 2010, 132, 3640–
3641.
[14] K. Shah, Y. Liu, C. Deirmengian, K. M. Shokat, Proc. Natl. Acad. Sci. USA
1997, 94, 3565–3570.
[15] S. T. Eblen, N. C. Kumar, K. Shah, M. J. Henderson, C. K. Watts, K. M.
Shokat, M. J. Weber, J. Biol. Chem. 2003, 278, 14926–14935.
[16] A. Clements, J. R. Rojas, R. C. Trievel, L. Wang, S. L. Berger, R. Marmor-
stein, EMBO J. 1999, 18, 3521–3532.
[17] M.-H. Kuo, J. Zhou, P. Jambeck, M. E. A. Churchill, C. D. Allis, Genes Dev.
1998, 12, 627–639.
[18] L. Wang, L. Liu, S. L. Berger, Genes Dev. 1998, 12, 640–653.
[19] W. H. H. Gunther, H. G. Mautner, J. Am. Chem. Soc. 1965, 87, 2708–2716.
[20] T. L. Miller, G. L. Rowley, C. J. Stewart, J. Am. Chem. Soc. 1966, 88, 2299–
2304.
[21] M. Shimizu, O. Nagase, Y. Hosokawa, H. Tagawa, Y. Yotsui, Chem. Pharm.
Bull. 1970, 18, 838–841.
Experimental Section
Compound syntheses: Syntheses of the target (1a and 1b) and
intermediate (2–6) compounds were carried out as described in
the Supporting Information. The early states of the synthesis were
conducted on a 0.5 mmol scale whilst the enzymatic conversion of
2-O-benzyl-4-O-phospho-d-pantethine to 3’’-O-benzyl-coenzyme A
was typically performed on a 20 mmol scale to generate ~1 mg of
the CoA analogue.
Protein expression and enzyme activity measurements: His-
tagged proteins of PPAT, DPCK and PCAF and its Q581A mutant
were purified by Ni2+ affinity column chromatography as detailed
in the Supporting Information.
PCAF HAT inhibition measurement: To investigate the inhibitory
effect, a fixed concentration of [2-14C]acetyl-CoA (10 mm) was used
with a varied amount of acetyl-3’’-O-benzyl-CoA (1b; 0–2 mm). Ra-
dioactivity was quantified by phosphorimage analysis by compari-
son to a known quantity of 14C-labelled bovine serum albumin
(10 mCimLÀ1, Cat. No. A-4844, Sigma). The IC50 value was deter-
mined by plotting the % enzyme activity against the log[acetyl-3’’-
O-benzyl-CoA] and fitting to a one-site competitive model by
using Prism (Graphpad.com). In all cases, background acetylation
(in the absence of enzyme) was subtracted from the total signal.
The activity was an average from triplicates of two different experi-
ments.
[22] J. F. A. Chase, B. Middleton, P. K. Tubbs, Biochem. Biophys. Res. Commun.
1966, 23, 208–213.
[23] E. J. Simon, D. Shemin, J. Am. Chem. Soc. 1953, 75, 2520.
[24] J. R. Stern, Methods Enzymol. 1955, 1, 559–566.
[25] J. F. A. Chase, P. K. Tubbs, Biochem. J. 1969, 111, 225–235.
[26] P. Rubenstein, R. Dryer, J. Biol. Chem. 1980, 255, 7858–7862.
[27] H. M. Miziorko, P. R. Kramer, J. A. Kulkoski, J. Biol. Chem. 1980, 255,
2842–2847.
[28] C. J. Stewart, T. Wieland, Liebigs Ann. Chem. 1978, 1, 57–65.
[29] M. Michenfelder, J. Rꢁtey, Angew. Chem. 1986, 98, 337–338; Angew.
Chem. Int. Ed. Engl. 1986, 25, 366–367.
Acknowledgements
[30] M. Shimizu, O. Nagase, S. Okada, Y. Abiko, T. Suzuki, Chem. Pharm. Bull.
1966, 14, 683–686.
[31] A. M. Michelson, Biochim. Biophys. Acta 1964, 93, 71–77.
[32] D. P. Martin, D. G. Drueckhammer, J. Am. Chem. Soc. 1992, 114, 7287–
7288.
[33] D. P. Martin, R. T. Bibart, D. G. Drueckhammer, J. Am. Chem. Soc. 1994,
116, 4660–4668.
[34] R. T. Bibart, K. W. Vogel, D. G. Drueckhammer, J. Org. Chem. 1999, 64,
2903–2909.
[35] K. W. Vogel, L. M. Stark, P. K. Mishra, W. Yang, D. G. Drueckhammer,
Bioorg. Med. Chem. 2000, 8, 2451–2460.
[36] E. Strauss, T. Begley, J. Biol. Chem. 2002, 277, 48205–48209.
[37] I. Nazi, K. P. Koteva, G. D. Wright, Anal. Biochem. 2004, 324, 100–105.
[38] J. G. Moffatt, H. G. Khorana, J. Am. Chem. Soc. 1959, 81, 1265–1265.
[39] M. van Wyk, E. Strauss, Chem. Commun. 2007, 398–400.
[40] R. C. Clough, S. R. Barnum, J. G. Jaworski, Anal. Biochem. 1989, 176, 82–
84.
This work was funded by BBSRC grants BBE0140891 and B15674
to E.S. and N.R.T. We also would like to thank Dr. Erick Strauss
(Department of Chemistry, Stellenbosch University, South Africa)
for providing the plasmids pESC106/CoaD pESC124/CoaE, and
Prof. Dale Drueckhammer (Department of Chemistry, State Uni-
versity of New York, Stony Brook) and Dr. Rob Layfield, (School of
Biomedical Sciences, University of Nottingham) for useful discus-
sions, and Sarentha Chetty for molecular graphics.
Keywords: acetyl-coenzyme A · DPCK · enzymes · PCAF ·
PPAT · transferases
[1] X.-J. Yang, V. V. Ogryzko, J. Nishikawa, B. H. Howard, Y. Nakatani, Nature
1996, 382, 319–324.
Received: May 14, 2010
[2] Z. Nagy, L. Tora, Oncogene 2007, 26, 5341–5347.
Published online on September 6, 2010
ChemBioChem 2010, 11, 2100 – 2103
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
2103